ENDODONTOE) LAND SNAILS FROM PACIFIC ISLANDS (Mollusca: Pulmonata: Sigmurethra) PartH Families Punctidae and Charopidae, Zoogeography Alan Solem Field Museum of Natural History Chicago, Illinois . UNlVLhSiTY OF ILLINOIS LIBRARY AT. URBANA -CHAMPAIGN ENDODONTOID LAND SNAILS FROM PACIFIC ISLANDS (Mollusca: Pulmonata: Sigmurethra) Part II Families Punctidae and Charopidae, Zoogeography Alan Solern Curator of Invertebrates Field Museum of Natural History Field Museum of Natural History Chicago, Illinois Submitted December 29, 1978 Published by: Field Museum of Natural History December 30, 1982 Mailed January 7, 1983 Library of Congress Catalog Card No.: 76-9516 ISSN 0015-0754 PRINTED IN THE UNITED STATES OF AMKRICA II CONTENTS LIST OF FIGURES V LIST OF TABLES VIII INTRODUCTION 1 Acknowledgements 2 PREVIOUS STUDIES 3 MATERIAL STUDIED 4 METHODS OF ANALYSIS 5 PATTERNS OF MORPHOLOGICAL VARIATION 8 Size and shape variations 8 Shell sculpture 9 Apertural barriers 15 Parietal barriers 19 Columellar barriers 21 Palatal barriers 21 Patterns of barrier variation 23 Summary of barrier variations 23 Summary of shell variations 23 Gross anatomy 23 Genital system 24 Ovotestis 24 Hermaphroditic duct 24 Talon and carrefour 25 Albumen gland 25 Prostate and uterus 25 Terminal male genitalia 25 Terminal female genitalia 30 Pallial complex 30 Digestive system 32 Free muscle system 34 Nervous system 34 External body features 34 Patterns of visceral hump reduction 36 Summary of anatomical variation 37 CHAROPID-ENDODONTID CONCHOLOGICAL COMPARISONS 38 HABITAT RANGE AND EXTINCTION 45 PHYLOGENY AND CLASSIFICATION 46 FAMILY CLASSIFICATION OF THE ENDODONTOIDS 47 PHYLOGENETIC PROCEDURES 48 GENERIC CLASSIFICATION 49 SYSTEMATIC REVIEW 50 List of the taxa 50 Geographic keys to the genera 51 Family Punctidae 53 Genus Punctum Morse, 1864 57 Family Charopidae 59 Subfamily Rotadiscinae 70 Genus Microcharopa, new genus 70 Subfamily Charopinae 72 Genus Discocharopa Iredale, 1913 74 Genus Sinployea, new genus 81 III IV Society Island Sinployea 85 Cook Island Sinployea 94 Samoan and Swains Island Sinployea 117 Western Polynesian and Micronesian Sinployea 133 Fijian Sinployea 145 Melanesian Sinployea 166 Genus Ba, new genus 172 Genus Maafu, new genus 174 Genus Lauopa, new genus 177 Genus Tuimalila, new genus 178 Genus Lagivala, new genus 183 Genus Vatusila, new genus 191 Genus Graeffedon, new genus 200 Subfamily Trukcharopinae, new subfamily 205 Genus Trukcharopa, new genus 207 Genus Kubaryiellus, new genus 210 Genus Russatus, new genus 214 Genus Roimontis, new genus 217 Genus Palikirus, new genus 219 Genus Jokajdon, new genus 222 Genus Palline, new genus 228 Subfamily Semperdoninae, new subfamily 235 Genus Semperdon, new genus 236 Genus Ladronellum, new genus 255 Genus Himeroconcha, new genus 259 Incertae sedis 266 ZOOGEOGRAPHY 268 Patterns of family and generic distributions 268 Patterns of diversity 275 Hawaiian Islands 275 Marquesas Islands 277 Tuamotu Archipelago 279 Mangareva, Gambier Islands 280 Rapa Island 282 Austral Islands 287 Society Islands 288 Cook Islands 289 Samoa 290 Tonga 292 Lau Archipelago 292 Fiji, Main Islands 295 Ellice Islands 295 Marshall Islands 295 Caroline Islands 297 Mariana Islands 297 Palau Islands 298 Summary of specific and generic diversity 301 Shell size and geography 305 Hypothesized origin and radiation of Pacific Island endodontoid snail fauna 307 Patterns of species diversity 310 Summary of zoogeography 315 SUMMARY 316 REFERENCES 317 APPENDIX: Anatomical terms 324 INDICES . 325 LIST OF FIGURES 1. Shell microsculpture inSinployea modicella 10 2. Shell microsculpture inSinployea peasei and Tuimalila pilsbryi 11 3. Shell microsculpture inPalline biakensis 12 4. Shell microsculpture in Microcharopa mimula 13 5. Details of shell sculpture inDiscocharopa aperta 16 6. Microstructure of largest palatal barrier inJokajdon callizonus 18 7. Surface of largest palatal barrier in Graeffedon graeffei 19 8. Structure of apertural barriers in Semperdon xyleborus 20 9. Anatomy ofCharopa coma 26 10. Interior of lower vas deferens and epiphallus in Charopa coma 27 11. Anatomy of Phenacohelix pilula 28 12. Radular teeth ofPunctum minutissimum 33 13. Radular teeth of Tuimalila pilsbryi 35 14. Radular teeth of Tuimalila pilsbryi 36 15. Mean shell height distribution in the Endodontidae and Charopidae 38 16. Mean shell diameter distribution in the Endodontidae and Charopidae 39 17. Mean height/diameter ratio distribution in the Endodontidae and Charopidae 39 18. Mean whorl count distribution in the Endodontidae and Charopidae 39 19. Mean diameter/umbilical width ratio distribution in the Endodontidae and Charopidae 40 20. Mean spire height/body whorl width ratio distribution in the Endodontidae and Charopidae 40 21. Mean ribs on body whorl distribution in the Endodontidae and Charopidae 40 22. Mean ribs/mm, on the body whorl distribution in the Endodontidae and Charopidae 41 23. Patterns of whorl width increment in selected Endodontidae and Charopidae 42 24. Partial anatomy ofPunctum minutissimum 54 25. Anatomy ofLaoma leimonias 55 26. Partial anatomy ofParalaoma aprica 57 27. Anatomy of "Paralaoma" coesa 58 28. Punctum polynesicum 60 29. Anatomy of Flammulina zebra 64 30. Anatomy ofAmphidoxa marmorella 65 31. Anatomy ofStephanoda binneyana 67 32. Anatomy of Pseudocharopa lidgbirdi 69 33. Microcharopa mimula 71 34. Anatomy of Discocharopa aperta 75 35. Parietal barrier variation inDiscocharopa aperta 76 36. Discocharopa aperta from the New Hebrides 80 37. Australian Discocharopa aperta 81 38. Sinployea modicella and S. tahitiensis 87 39. Anatomy of Tahitian and Moorean Sinployea 90 40. Sinployea lamellicosta 91 41. Sinployea montana andS. neglecta 93 42. Anatomy of Society and Cook Island Sinployea 96 43. Anatomy of Sinployea avanaensis 97 44. Sinployea andrewi andS. atiensis 100 45. Sinployea peasei and S. avanaensis 103 46. Sinployea proximo and S. planospira 107 47. Sinployea decorticata andS. rudis 109 48. Sinployea harveyensis and S. youngi 113 49. Sinployea canalis 115 50. Sinployea otareae and S. tenuicostata 116 V VI 51. Sinployea clista andS. clausa 120 52. Sinployea aunuuana andS. intermedia 123 53. Anatomy of Sinployea aunuuana, S. clista, and S. intermedia 124 54. Sinployea allecta allecta and S. a. tauensis 126 55. Anatomy of Sinployea allecta allecta 127 56. Sinployea complementaria 130 57. Anatomy of the Samoan Sinployea complementaria 132 58. Sinployea vicaria vicaria and S. v. paucicosta 136 59. Shell sculpture details of Sinployea vicaria vicaria 137 60. Anatomy of the Tongan Sinployea vicaria vicaria 138 61. Sinployea pseudovicaria andS. rotumana 139 62. Sinployea ellicensis ellicensis and S. e. nukulaelaeana 141 63. Sinployea kusaieana and Trukcharopa trukana 144 64. Sinployea recursa andS. angularis 148 65. Sinployea princei andS. irregularis 151 66. Sinployea inermis inermis, S. i. meridionalis, and S. i. lakembana 156 67. Anatomy of Fijian, Caroline Island, and New Hebridean Sinployea 158 68. Sinployea lauensis andS. navutuensis 160 69. Sinployea godeffroyana andS. adposita 162 70. Sinployea euryomphala 168 71. Sinployea solomonensis 169 72. Sinployea kuntzi 170 73. Sinployea novopommerana and S. descendens 171 74. Ba humbugi 173 75. Anatomy ofBa humbugi 175 76. Maafu thaumasius 176 77. Lauopa mbalavuana 178 78. Tuimalila pilsbryi and T. infundibulus 180 79. Anatomy of the Tongan Tuimalila pilsbryi 182 80. Lagivala vivus andL. minusculus 186 81. Lagivala microglyphis andL. macroglyphis 188 82. Vatusila kondoi and V. nayauana 193 83. Vatusila tongensis and V. niueana 197 84. Partial anatomy of Vatusila tongensis 198 85. Vatusila uaitupuensis and V. eniwetokensis 199 86. Graeffedon graeffei 201 87. Anatomy of the Samoan Graeffedon graeffei 203 88. Graeffedon savaiiensis and G. pricei 204 89. Kubaryiellus kubaryi andRussatus nigrescens 213 90. Anatomy ofRussatus, Kubaryiellus, and Trukcharopa 216 91. Roimontis tolotomensis andPalikirus cosmetus 218 92. Jokajdon callizonus and J. tumidulus 224 93. Anatomy of Jokajdon andPalline 227 94. Palline notera notera and P. n. palauana 230 95. Palline notera gianda andP. micramyla 233 96. Palline biakensis 235 97. Semperdon uncatus 239 98. Semperdon kororensis andS. xyleborus 242 99. Semperdon heptaptychius andS. rotanus 246 100 Anatomy of Semperdon 249 101. Ladronellum mariannarum 256 102. Anatomy of Ladronellum mariannarum 258 103. Anatomy ofHimeroconcha 261 104. Himeroconcha rotula and//, lamlanensis 262 105. Himeroconcha quadrasi and//, fusca 264 106. Distribution of Pacific Island endodontoid families 269 107. Distribution of most generalized and most specialized endodontid genera 270 108. Ranges of Lagivala and Vatusila 271 109. Range of Mautodontha and areas where umbilical brood chambers evolved 272 110. Range of Sinployea . 273 VII 111. Ranges ofGraeffedon, Palline, and Semperdon 274 112. Collecting localities on smaller islets of Mangareva, Gambier Islands 280 113. Collecting localities on Mangareva Islet, Mangareva, Gambier Islands 281 114. Collecting localities on Rapa Island 283 115. Major geographic features on Rurutu, Austral Islands 287 116. Collecting localities on Tubuai, Austral Islands 288 117. Collecting localities on Raivavae, Austral Islands 289 118. Collecting localities on Rarotonga, Cook Islands 290 119. Collecting localities on Upolu, Samoa 291 120. Collecting localities on Savai'i, Samoa 291 121. Main islands of Tonga 292 122. Collecting localities on Vava'u, Tonga 293 123. Collecting localities on Tongatapu, Tonga 294 124. Collecting localities on Eua, Tonga 295 125. Islands of the Lau Archipelago 296 126. Collecting localities on Kusaie, Caroline Islands 298 127. Collecting localities on Ponape, Caroline Islands 298 128. Islets of the Truk Group, Caroline Islands 298 129. Collecting localities on Moen Islet, Truk, Caroline Islands 298 130. Collecting localities on Tol Islet, Truk, Caroline Islands 299 131. Collecting localities on Dublon Islet, Truk, Caroline Islands 299 132. Collecting localities on Guam, Mariana Islands 299 133. Islands of the Palau Group 300 134. Collecting localities on Peleliu, Palau Group 300 135. Collecting localities on Koror and Auluptagel, Palau Group 300 136. Collection station on Angaur, Palau Group 301 137. Species diversity in island groups for the Endodontidae and Charopidae 303 138. Generic diversity in island groups for the Endodontidae and Charopidae 304 139. Median mean and mean ranges of adult shell diameter in island groups for the Endodontidae and Charopidae 306 140. Species-area relationships for total land snail faunas of selected Pacific Islands 312 141. Species-area relationships for selected land snail families on Polynesian and Micronesian islands 313 142. Species diversity in islands grouped by size 313 143. Species diversity in islands grouped by elevation 313 LIST OF TABLES I. Range of variation in the Pacific Island Charopidae 8 II. Body whorl contour in Endodontidae and Charopidae 8 III. Shell diameter and rib spacing in the Charopidae 14 IV. Degree of apertural narrowing by barriers 18 V. Percentage distribution of columellar barrier numbers 21 VI. Percentage distribution of palatal barrier numbers 22 VII. Palatal barrier numbers in constricted aperture Charopidae 22 VIII. Phyletic representation of dissected taxa 24 IX. Relative length of kidney lobes in the Pacific Island Charopidae 31 X. Conchological comparisons of Endodontidae and Charopidae from Pacific Islands 38 XI. Species numbers of Endodontoid land snails from selected geographic areas 53 XII. Range of variation in Punctum, Discocharopa, and Microcharopa 59 XIII. Geographic variation in Discocharopa aperta 79 XIV. Ratio of penis length to shell diameter in Sinployea 83 XV. Mean species measurements in Sinployea 83 XVI. Range of variation in Society Island Sinployea 85 XVII. Local variation in Society Island Sinployea, 1 89 XVIII. Local variation in Society Island Sinployea, II 95 XIX. Range of variation in Cook Island Sinployea 98 XX. Local variation in Cook Island Sinployea, 1 101 XXI. Local variation in Sinployea peasei 104 XXII. Local variation in Cook Island Sinployea, II Ill XXIII. Local variation in Cook Island Sinployea, III 112 XXIV. Range of variation in Samoan and Swains Island Sinployea 118 XXV. Local variation in Samoan Sinployea, 1 121 XXVI. Local variation in Samoan Sinployea, II 128 XXVII. Range of variation in Western Polynesian and Micronesian Sinployea 133 XXVIII. Local variation in Sinployea vicaria 135 XXIX. Local variation in Western Polynesian Sinployea 142 XXX. Range of variation in Fijian Sinployea 146 XXXI. Local variation in Fijian, Melanesian, and Micronesian Sinployea 152 XXXII. D/U ratio frequency in Sinployea inermis subspecies 153 XXXIII. Ribs/mm, frequency in Sinployea inermis subspecies 153 XXXIV. Rib frequency in Sinployea inermis subspecies 153 XXXV. Local variation in Sinployea inermis, 1 154 XXXVI. Local variation in Sinployea inermis, II 155 XXXVII. Local variation in Sinployea adposita andS. irregularis 164 XXXVIII. Range of variation in Melanesian Sinployea 167 XXXIX. Range of variation in Fijian and Tongan endemic genera 181 XL. Range of variation inLagivala 184 XLI. Local variation inLagivala demani 191 XLII. Range of variation in Vatusila 194 XLIII. Range of variation in Graeffedon 202 XLIV. Sympatry of Charopidae on Ponape, Caroline Islands 208 XLV. Range of variation in Micronesian Charopinae, 1 209 XLVI. Local variation in Trukcharopa 211 XLVII. Age distribution in Trukcharopa trukana 211 XLVIII. Local variation in Kubaryiellus andRussatus 215 XLIX. Local variation inPalikirus andJokajdon 220 L. Range of variation in Micronesian Charopinae, II 223 VIII IX LI. Sculptural variation inJokajdon tumidulus 226 LII. Local variation inPalline 232 LIII. Range of variation in Semperdoninae 237 LIV. Local variation in Palau Island Semperdon 240 LV. Summary of barrier variation in Mariana Islands Semperdon 241 LVI. Local variation in Semperdon rotanus 248 LVII. Barrier variation in Semperdon rotanus 248 LVIII. Local variation in types and large form of Semperdon heptaptychius 252 LIX. Local variation in Semperdon heptaptychius 253 LX. Sculpture variation in Semperdon heptaptychius 254 LXI. Frequency distribution of diameter in populations of Semperdon heptaptychius 254 LXII. Frequency distribution of ribs in populations of Semperdon heptaptychius 254 LXIII. Frequency distribution of ribs/mm, in populations of Semperdon heptaptychius 254 LXIV. Barrier variation in Semperdon heptaptychius 255 LXV. Local variation in Ladronellum and Himeroconcha 257 LXVI. Generic distribution of Hawaiian Endodontidae and Punctidae 276 LXVII. Geographic distribution of species level taxa in Hawaiian Endodontidae 276 LXVIII. Diversity of Hawaiian land snails 276 LXIX. Land snail fauna of Rapa Island 284 LXX. Ages of Pacific Islands 307 LXXI. Pacific Island land snail families with comprehensive distributional data available 311 LXXII. Land snail faunas of Lord Howe, Upolu, and Viti Levu 312 LXXIII. Calculated and observed maximum density of land snail species 312 LXXIV. Correlation between island size, elevation, and number of land snail species 314 LXXV. Land snail diversity on islands of same size but different distances from Indonesian-Melane- sian core 314 LXXVI. Diversity of land snails on small, isolated islands 314 INTRODUCTION This is the second of two monographs revising the endodontoid land snails of Polynesia, Micronesia, and Fiji. Part I, covering the older and larger family En- dodontidae, reviewed 185 species-level taxa (Solem, 1976b). Subsequently (Solem, 1977a) I described Pro- toendodonta laddi and Cookeconcha antiquus from Late Pleistocene strata in core drillings on Midway Atoll and have three new subfossil Pseudolibera from Makatea (Solem, in preparation B). Together with the 98 species-level taxa belonging to the Families Punc- tidae and Charopidae diagnosed below, a total of 290 species-level taxa have been reviewed during this pro- ject. Describing the extensive Hawaiian Island monophyletic radiation of up to 290 unnamed, now mostly extinct, endodontid taxa preserved in the Ber- nice P. Bishop Museum is a task left to others. The basic descriptive work essentially was com- pleted in 1969, based on research supported by Na- tional Science Foundation grants G-16419, GB-3384, and GB-6779. This has been supplemented by modest materials received subsequently. As a by-product of the basic monographs, papers on a number of prob- lematic groups (Solem, 1964a-b, 1968b, 1969b, 1972b, 1973c-d), extralimital endodontoids (Solem, 1958b, 1961, 1968c, 1970a, c, 1975, 1976a, 1977c, in prepara- tion A), biogeography (Solem, 1958a, 1959a, 1968a, 1969c, 1973c, 1979a-b, 1981), endodontoid biology (Solem, 1969a, d, 1974), and fossil land snails have been prepared (Solem, 1977a, in preparation B; Solem & Yochelson, 1979). In the early 1970s, the new availability of scan- ning electron microscopy (hereafter SEM) added a major new dimension to this project. A series of reports (Solem, 1970b, 1972a, c, 1973a-b, f, 1976a, 1977a-b) explored its potential, and extensive additional data has been incorporated into the major reviews. The first monograph was submitted for publication in De- cember, 1973, and was issued in due course (Solem, 1976b). Finishing Part II has involved making very dif- ficult decisions as to where and when to terminate additional information seeking. The Endodontidae are restricted to the Pacific Islands and thus this mono- graph was easily circumscribed. Both the Punctidae and Charopidae have their main centers of diversity elsewhere, have been liberally supplied with generic and family-level names based on shell features and hunches, frequently have been rather poorly described and often never illustrated (particularly Iredale, 1913, 1933, 1937a-c, 1939, 1941a-c, 1942, 1944, 1945), and essentially are unknown anatomically except for the data supplied by Hutton (1884a), Suter (see Refer- ences), and Climo (1969a-b, 1970, 1971a-b, 1978) on New Zealand taxa and contributions on South African taxa summarized by Connolly (1939). I have dissected and partially illustrated more than 75 extralimital species in addition to those included below. This attempt to work out extralimital relationships and to assign subfamily and generic names is only a start toward producing an overall phylogeny and classifica- tion. The limits finally adopted in Part II were: (1) to include all extralimital species known to belong in genera endemic to Polynesia, Micronesia, and Fiji; and (2) to include anatomical data on extralimital taxa only when needed to establish the subfamily units used in the systematic review and biogeography. Thus, 10 of the 98 taxa discussed here are found only in such areas as the New Hebrides, Solomon Islands, Bismarck Ar- chipelago, West Irian, and Indonesia (six species of Sinployea, three species of Lagivala, and Palline biakensis), whereas another species, Discocharopa aperta (Mollendorff, 1888), occurs sporadically in the area covered (Fiji, Samoa, Austral and Society Islands) but has its main distribution from the Philippines to Australia. Data are presented on type genera or species of several nominate subfamilies from diverse areas (New Zealand, Juan Fernandez), but no attempt is made to provide an overall classification of the Charopidae in this monograph. There still are too many gaps in our knowledge of anatomy and shell structure. By the early 1970s there still remained a few prob- lem taxa. Microcharopa and Discocharopa showed many differences from the other Pacific Island taxa, but I had not been able to determine their affinities because only empty shells were available. In addition, there was the problem of Stenopylis, a second Philip- pine Islands to Australia taxon. Both Stenopylis and Discocharopa were known to occur in Central Aus- tralia. Planned fieldwork in Western and Central Aus- tralia in early 1974 hopefully would yield live mate- rial. Early dry season collecting near Alice Springs produced only dead shells, but museum records sug- gested that both genera had a wide distribution in North Australia. The probability of subsequently ob- taining live materials seemed excellent. In addition, the possibility of field survey work in Micronesia for the Office of Endangered Species arose, which could have greatly expanded the anatomical coverage of the SOLEM: ENDODONTOID LAND SNAILS Trukcharopinae and Semperdoninae. This prospect did not materialize, but in 1976 and 1977 additional field and dissection work in Australia permitted allocating Discocharopa, Stenopylis, and Microcharopa to family units. The writing of introductory sections and the biogeographic review, assigning table and plate num- bers, and final editing were completed in 1978. It is with a mixture of delight and regret that I conclude this systematic project delight at the new information and ideas produced, regret that the hand of man has so altered the Pacific Island environment that future work on most of these taxa and testing of the new ideas presented has been rendered impossible by extinctions. The charopids apparently have been less affected than the Endodontidae, but even with these it is probable that many already are a vanished group of organisms. Concluding this monograph is a detailed biogeo- graphic review of the endodontoid land snails. This project spanned the time from the dominance of fixed continents as a biogeographic axiom to the current dogma of wandering continents. Land snails have not been versed in biogeographic controversies, and their patterns of distribution continue to fit neither model exclusively. The endodontoid land snails were the most spe- ciose land snail group on the Pacific Islands (Solem, 1976b, pp. 1-2). They still are the most speciose group in Australia and New Zealand (table XI), areas in which they seem to be flourishing despite human ac- tivities. The data and ideas presented in these volumes hopefully will inspire more investigations of these taxa. ACKNOWLEDGEMENTS In addition to the many individuals listed by Solem (1976b, pp. 5-6), the following people have con- tributed significantly to the success of this project. For access to specimens in collections, suggestions as to collecting localities, loan of materials, and many help- ful ideas, I am indebted to Dr. Frank Climo, National Museum of Natural History, Wellington, New Zealand; Drs. Edmund Gittenberger and A. C. van Bruggen, Rijksmuseum van Natuurlijke Historie, Leiden; Dr. Jackie van Goethem, Institut Royal des Sciences Naturelles de Belgique, Brussels; Dr. Simon Tillier, Museum National d'histoire Naturelle, Paris; Dr. Peter Mordan, British Museum (Natural History), London; Dr. Barry Wilson, Dr. Fred Wells, Ms. Shirley Slack-Smith, Mr. George Kendrick, Western Austra- lian Museum, Perth; Dr. Winston Ponder, Australian Museum, Sydney; and Mr. and Mrs. Fred Aslin, Mt. Gambier, South Australia. For assistance in the field, or supplementary fieldwork by, I continue to be deeply in debt to Mr. Laurie Price, Kaitaia, New Zealand, Field Associate of Field Museum of Natural History. Mr. Carl Christen- sen, University of Arizona; Dr. Barry Wilson, Shirley Slack-Smith, George Kendrick, Clayton Bryce, and many other staff members of the Western Australian Museum; Fred and Jan Aslin, Mt. Gambier, South Australia; Dr. Brian J. Smith, National Museum of Victoria, Melbourne; Dr. Winston Ponder, Australian Museum, Sydney; and Mr. Ron Kershaw, Launceston, Tasmania, have assisted in Australian fieldwork or have provided critical help with collecting localities. Charts and figures in the introductory sections have been prepared new or reworked by Ms. Elizabeth Liebman, Illustrator, Division of Invertebrates, Field Museum of Natural History. Invaluable help with mounting and labeling was provided by Mrs. Dorothy Karall, Associate, Division of Invertebrates. Illus- trations for the systematic review were done primarily by the several artists cited by Solem (1976b, p. 6). The credit for the individual figures is indicated by the fol- lowing abbreviations: Ms. Nelva Bonucchi (NB); Mrs. Jane Calvin (JC); Mrs. Carole W. Christman (CW); Mr. Sam Grove (SG); Mrs. Lynda Hanke (LH); Mr. Sander Heilig (SH); Dr. Harold J. Walter (HJW); Mr. E. John Pfiffner (EJP); Mrs. Patricia Rill Smiley (PS); Ms. Marcia Oddi (MO); and the late Mrs. Margaret Anne Moran McKibben (MM). In particular, the work by the late Margaret Anne Moran McKibben was critical to the success of this project. Tabular material in the systematic review was typed by Mrs. Sandra Rendleman and Ms. Jayne Freshour. Tabular material in the introductory sec- tions was typed by Mrs. Valerie Connor-Jackson, Sec- retary, Division of Invertebrates. Grateful acknowledgement is made to the Na- tional Science Foundation, whose generous support through grants G- 16419, GB-3384, GB-6779, and DEB 75-20113 in support of fieldwork, illustration prepara- tion, and museum study of types has made this study possible. The establishment of a scanning electron microscope laboratory at Field Museum of Natural History was funded by National Science Foundation grant BMS 72-02149. Publication of Part I (Solem, 1976b) was in large part made possible by National Science Foundation grant DEB 75-14048. Much help with final editing and complete typing of the manuscript was given by Ms. Sharon Bacoyanis and Mrs. Valerie Connor-Jackson. Their efficient in- volvement in every phase of final preparation was crit- ical. Proofreading and indexing was greatly aided by Valerie Connor-Jackson. As with Part I, the years of labor by the late C. Montague Cooke, and by Yoshio Kondo, plus the facilities provided by the Bernice P. Bishop Museum during 1961-1962 and their long loan of needed mate- rials were instrumental to the initiation and comple- tion of this project. Dr. Cooke had not reviewed any of the taxa covered in this report, so all errors are my responsibility. PREVIOUS STUDIES A more extensive historical review regarding the Pacific Island taxa was given in Solem (1976b, pp. 7-8) and should be consulted for background comments. Previously described charopid taxa date from the studies of Ferussac (1840), Hombron & Jacquinot (1841), Mousson (1865, 1869, 1870, 1871), Cox (1870), Pease (1870), Garrett (1872, 1881, 1884, 1887a-c), Liardet (1876), Tapparone-Canefri (1883), Mollendorff (1888, 1900), Beddome (1889), Quadras & Mollendorff (1894), Smith (1897), Clapp (1923), Rensch (1937), Dell (1955), Ladd (1958, 1968), and Solem (1959a, 1960). Of the 45 previously described species-level taxa, 11 were named by Garrett, six by Mollendorff, five by Mousson, four by Quadras & Mollendorff, and four extralimital species by Rensch. The remaining 15 species are scat- tered among the other authors. The dates of species descriptions were summarized in Solem (1976b, p. 3, table III). The cluster of 1890 and 1900s taxa in the Charopidae reflects the activities of Mollendorff and his field associate, J. F. Quadras. The 10 taxa described since 1920 include the highly significant fossils from Eniwetok, Vatusila eniwetokensis (Ladd, 1958), and Funafuti, Lagivala davidi (Ladd, 1968), plus basically extralimital taxa (Solomon Islands, New Hebrides, Bismarck Archipelago) that happen to belong to gen- era also represented on the Pacific Islands. As with the Endodontidae, the greatest period of descriptive activ- ity was in the 1870s and 1880s, reflecting the publica- tions of Mousson and Garrett. No data had been recorded in the literature con- cerning the anatomy of the Pacific Island taxa except for a few notes on the tail, jaw, and radula ofRussatus nigrescens (Mollendorff, 1900, p. 108, figs, 1-3) and the radula and penis complex exterior of Sinployea euryomphala (Solem, 1959a, pi. 6, figs. 7-8). Because the main center of the charopid radiation is extralimital to the Pacific Islands, extensive refer- ence has had to be made to taxa from other parts of the world. New Zealand, Australian, and New Caledonian taxa in particular have yielded important comparative data. Rather than detail the many reports here, cita- tions are given under the organs or species referred to at that place in the text. Here it is necessary only to point out that these are diverse faunas with a long record of study. The literature is highly fragmentary and widely scattered, but fortunately there are sum- mary references that enable relatively quick access to the literature. For New Zealand, the classic mono- graph of Suter (1913) and the checklist of Powell (1976) summarize the work done prior to the highly signifi- cant studies of Climo (1969a-b, 1970, 1971a-b, 1978). For Australia, the checklist of Iredale (1937a-c) and his faunistic reviews of nearby areas as well as Aus- tralia (Iredale, 1939, 1941a-c, 1942, 1944, 1945) pro- vide essential but very difficult to use reports. Franc (1957) and Solem (1961) supplement the excellent re- view of New Caledonian land mollusks given by Crosse (1894). Early observations on the anatomy and radular structure of some Australian (Hedley, 1891, 1893) and New Zealand (Hutton, 1884a, Suter, 1890, 1891a-b, 1892a-c, 1893a-d, 1894a-d, 1901, 1903, 1913) charopids were accurate within the limits of optical observation and the state of knowledge concerning molluscan anatomy extant at that time, but have little value today in deciding generic and family limits. Even more than for the Endodontidae, the classifi- cation of the charopids used previously was based on form "pigeon holes," and the assignments made to gen- era were based upon superficial conchological simi- larities. Of the 21 genera reviewed below, only two were named previously, and both of them, Punctum Morse, 1864, and Discocharopa Iredale, 1913, are gen- era whose main distributions lie outside the Pacific Island area. All the remaining species belong to unde- scribed, endemic taxa, mostly not yet known from other than the fringing islands and Pacific Island area proper. MATERIAL STUDIED The 98 taxa formally described or alluded to as probably representing new taxa were represented by about 7,500 specimens. Since the summary presented by Solem (1976b, p. 9), a number of charopids from the Lau Archipelago of Fiji collected by Price in 1970 and Cernohorsky in 1977, Society Islands by Solem in 1977, Western Australia by Solem, L. Price, and Carl Chris- tensen in 1976 and 1977, and near Alice Springs in the Northern Territory in 1974 and 1977 have become available. Data from them has been incorporated into the systematic review. No attempt at recalculating the data in Solem (1976b, pp. 9-10, tables IV- V) has been made because the numbers added are small in com- parison to the previously listed materials. Throughout the text, the following abbreviations are used to indicate the repository of the cited speci- mens: AIM Auckland Institute and Museum, Auckland AMS Australian Museum, Sydney ANSP Academy of Natural Sciences, Philadelphia BMNH British Museum (Natural History), London BPBM Bernice P. Bishop Museum, Honolulu DMW National Museum of Natural History, Wellington FMNH Field Museum of Natural History, Chicago IRB Institut Royal des Sciences Naturelles de Belgique, Brussels JDCP Collection of Journal de Conchyliologie, Paris MCZ Museum of Comparative Zoology, Harvard Univer- sity, Cambridge MSNG Museo Civico di Storia Naturale "Giacomo Doria," Genova NHM Naturhistorisches Museum, Basel NMWC National Museum of Wales, Cardiff RNHL Rijksmuseum van Natuurlijke Historic, Leiden RSM Royal Scottish Museum, Edinburgh SAM South Australian Museum, Adelaide SMF Natur-Museum Senckenberg, Frankfurt-a.-M. UMMZ University of Michigan, Museum of Zoology, Ann Arbor USNM National Museum of Natural History, Smithsonian Institution, Washington, D.C. WAM Western Australian Museum, Perth ZMA Zoologisch Museum, Amsterdam ZMB Zoologisches Museum der Humboldt-Universitat, Berlin Zurich Zoologisches Institut der Universitat, Zurich Most of the material in the BPBM resulted from the historical activities of Andrew Garrett plus speci- mens obtained during three major expeditions in the 1930s. The Mangarevan Expedition from April 15 through October 28, 1934 (Cooke, 1935), Micronesian Expedition from December 8, 1935, to June 10, 1936 (Gregory, 1936, p. 40), and the Henry G. Lapham Ex- pedition to Fiji from June 27 through September 28, 1938 (Buck, 1939, pp. 29-30), contributed major mate- rials for this study. Additional collections made in Fiji by Harry G. Ladd, plus miscellaneous samples taken by a variety of BPBM staff members, combined with the above to provide the bulk of material analyzed. Fieldwork sponsored by FMNH and National Sci- ence Foundation in Rarotonga (1964 and 1965), West- ern Samoa (1965), Fiji (1965 and 1970), Tonga (1965- 1966), New Caledonia (1967), and New Hebrides (1972) by Laurie Price and/or Alan Solem was designed to sample areas that were underrepresented in the BPBM collections. Fieldwork in New Zealand (1962, 1965) and various parts of Australia (1962, 1965, 1973, 1974, 1976-77) provided critical comparative materials that enabled assigning subfamily names and indicating possible phylogenetic trends within the Arionacea. Only a few of these dissections are illustrated or cited here because this report is deliberately restricted to the Pacific Islands, except where a genus extends extralim- itally. Special features regarding the methods of collect- ing and specimen storage used by the BPBM staff were reviewed by Solem (1976b, p. 10) and are not repeated here, although applying in full to the studied materials used in this report. METHODS OF ANALYSIS The basic data on distribution of size and shape variation within a population, how to recognize an adult shell, the limits of measurement reliability, and definitions of the standard measurements used were fully explained in Solem (1976b, pp. 11-15). It is necessary to emphasize some data concerning sample bias because rather significant mean size differences among museum lots of previously described species are sometimes ignored. Prime examples of this are in such Rarotongan taxa as Sinployea proxima (Garrett, 1872) (table XX), S. rudis (Garrett, 1872) (table XXIII), S. harveyensis (Garrett, 1872) (table XXII), and the Funafuti Atoll S. ellicensis ellicensis (table XXIX). The pattern of trading samples from larger series to other museums or collectors still is common practice among conchologists. Larger speci- mens will tend to be retained by the original source, with smaller examples being dispersed widely. The possible extent of such dispersal is perhaps best illus- trated by the endodontid land snail species Libera tumuloides (Garrett, 1872). Although "over 300 speci- mens" were collected originally, only 72 could be lo- cated for this study, indicating dispersion and loss or at least absence from major museum collections of more than 228 examples in less than 100 years (Solem, 1976b, p. 430). In contrast, the endodontid Mauto- dontha (M.)zebrina (Garrett, 1874) was known from "a half dozen specimens," all of which were located (Solem, 1976b, p. 162). Hence the reliability of mean measurements made on rare species is apt to be much greater than the means for common species where trading bias probably has distorted the data base. The exact extent of this distortion is unknown because the original measurements were generally of only a single specimen and were cited to the nearest millimeter. It must be recognized that sizes cited here for pre-1900 species may be larger than the means of the popula- tions from which they came, but how much larger is unknown. Data from the BPBM and FMNH collections in- volved measuring field samples that had not been di- luted by trading activities. Thus, these results are di- rectly comparable with each other. For each adult specimen, shell height, diameter, whorl count, umbilical width, number of radial ribs on the body whorl, number of apical spiral cords, and any information concerning number and position of aper- tural barriers were recorded. The only difference in data recording from that used in Solem (1976b) is the consistent counting of body whorl ribs and apical cords. The lack of apertural barriers in many charopids suggested greater use of sculptural features, and thus body whorl ribs were counted wherever possible. Solem (1976b, p. 14, fig. 5; p. 43, fig. 32; p. 52, fig. 36) defined each of these parameters except for the apical spiral cord count because the Endodontidae normally lacks this type of major sculpture. This count was taken on a suture-to-suture transect just before the end of the apex, provided that portion of the shell was unworn. At the start of the apex the exposed portion of the whorl profile is narrower than at the end, producing a sig- nificantly lower spiral ridge count. In Microcharopa mimula (fig. 4a), for example, the apex tip shows 15 ribs, and the apex end, 20 ribs. Usually, if a count could not be made near or at the usual position, the apex was sufficiently worn so that no count could be obtained. Apical cord count could not be tallied on many speci- mens. It was, in general, subject only to minor error in tallying where countable. Height/Diameter (hereafter H/D) ratios, Diameter/ Umbilical Width (hereafter D/U) ratios, and Ribs/mm, of body whorl circumference were calculated, the latter by the formula: rib count on body whorl Ribs/mm. = TT x shell diameter in mm. The validity of this index was discussed by Solem (1976b, pp. 42-43). Means, ranges, and standard errors of the means were computed for each parameter except whorl count for each population. Some of these results are pre- sented in tables headed "Local Variation in ... ." The total variation within each species or subspecies is summarized in tables headed "Range of Variation in . . . ." Means are given, but no standard errors of the mean were calculated because of the wide temporal span (1820 to 1977) of the materials represented. At a later stage in this review, from eight to 12 examples of each species, where sufficient materials were available, were measured as to spire elevation and body whorl width, with a suitable ratio (SP/BWW) calculated and included in the "Range of Varia- tion. . . ." tables. This gives an index of actual spire elevation undistorted by differences in the degree of body whorl descension behind the lip (see Solem, 1976b, pp. 24-26). At the same time these additional measurements were made, the pattern of the micro- sculpture on the early portion of the body whorl (see Solem, 1976b, p. 42, fig. 32b) and the pattern of rib spacing were determined and added to the draft diag- noses and descriptions. SOLEM: ENDODONTOID LAND SNAILS Dissections of as many taxa as possible were car- ried out between 1961 and 1977, with several rechecks of structure as my understanding of the variation pat- terns increased. Because only fragmentary preserved animals were available for many species (see Solem, 1976b, p. 19), often only data on the terminal portions of the genitalia could be presented. Where possible, usually both the entire genitalia and the penis complex interior plus the pallial region were illustrated. Data on body color, tail structure, muscle system, and radular structures were obtained, but have not been illustrated because of lesser information content for classification and phylogeny. Unfortunately, the bulk of the dissec- tion, illustration, and descriptive work was done prior to the availability of scanning electron microscopy as a research tool. Hence only a modest number of SEM observations are included in this study, although ex- tensive use has been made of the SEM in working with materials from extralimital areas. The data base consists of many shell observations and measurements, data from the pallial region, genitalia, radula, and external body features. In the previous report, some use was made of computer- generated phylogenies (Solem, 1976b, pp. 114117, figs. 59-61). No such attempts were made in studying the Charopidae, because major features such as the apertural barriers obviously arose independently in several lineages (Solem, 1973b, p. 305), there were sharp anatomical differences between major groups with highly similar shell features, and the resulting convergences would utterly confuse any phenetic pro- gram unless de post facto elimination of convergences was undertaken. The benefits of the computer model- ing outlined by Solem (1976b, pp. 114-117) were applied to this analysis without the work of program- ming and running various combinations. Mean mea- surement data was key-punched and numerous print- outs prepared, but this was essentially an aid to the phylogenetic analyses. Cladistic analysis was not at- tempted, because there are still numerous gaps in knowledge of the extralimital taxa. Outgroup compari- sons would have been sketchy or incomplete at best, since the large charopid and punctid faunas of New Zealand, Australia, New Caledonia, and associated is- lands occur under conditions of complex sympatry, with many interspecific adjustments in penial chamber features that make interpreting relationships exceed- ingly difficult. The result is that the proposed phylogeny has been arrived at by phylogenetic methodology after looking at a large number of characters in several organ sys- tems and interpreting these according to the basic three-tiered pattern of evolutionary events outlined in Solem (1978b). I have used a pragmatic rather than an ideologic approach. The problem of how to determine species limits when allopatric island populations are involved is quite difficult (Solem, 1976b, pp. 15-17). Sympatric congeneric species pairs or trios are invaluable in pre- dicting the probability that two allopatric populations are or are not reproductively compatible. Since the Charopidae, as do the Endodontidae, use the surface features of the penis chamber and lower female tract for species recognition when two closely related species are sympatric, there usually are demonstrable dif- ferences in this region of the reproductive tract. Corre- lated with this will be minor to prominent shifts in conchological structures. By determining the "mini- mum conchological difference" between sympatric con- generic species pairs, a pragmatic standard is available against which to measure the degree of conchological difference between allopatric populations. If they equal or exceed the "minimum conchological difference," they are presumed to have diverged genetically to the level of species. If they show slight differences that are below the "minimum conchological difference," they are judged to be conspecific or at most subspecifically separable. Under conditions of allopatry there are few to no selective pressures for changes in the species rec- ognition features of the genitalia. Indeed, there proba- bly is rather strong selective pressure for maintaining a conservative pattern of species recognition features because departure from the pattern might be a mating disadvantage. This methodology does not presume genetic link- age between conchological differences and species rec- ognition features. Quite the contrary pattern is hypothesized. The conchological differences are linked to physiological or behavioral traits that have been selected for by environmental pressures, producing eventual basic genetic incompatibility. Upon becoming microsympatric with a congener, strong selective pres- sure for species recognition changes would exist, if, as seems probable, nonmultiple mating is a standard pat- tern for the charopids. Although we have no data on the frequency of mating, the absence of any evidence of more than one set of transferred sperm during dissec- tion of several hundred individuals suggests that lim- ited mating, rather than multiple mating, is more probable. There are far fewer clear examples of congeneric sympatry with adequate materials for dissection in the Charopidae than in the Endodontidae (see Solem, 1976b, pp. 80-83, tables LII-LIV). Partly this reflects the difference in the distribution and probable age within the area. Partly this reflects a greater vertical range in habitat for the Charopidae, with tree trunk to semiarboreal habitat available even at dryer eleva- tions for the charopids as contrasted with the terres- trially limited endodontids. Partly it is a matter of the Charopidae, except for Sinployea, showing rather low speciation but high generic diversity within the Pacific Island area. The one island on which an apparent mul- titude of species in one genus existed, Rarotonga, ap- parently has nine of 11 charopid species extinct. The example that comes closest to being sympatric and showing minimal conchological differences in- volves Sinployea peasei and S. avanaensis on Raro- METHODS OF ANALYSIS tonga. The shell of S. avanaensis (fig. 45d-f) has more than twice the number of radial ribs on the body whorl, a distinctly narrower umbilicus, and averages about one-quarter whorl less, although it is nearly identical in size and shape. Anatomically, the penis of S. avanaensis is 1.41.6 mm. long, compared with 1.1-1.3 mm. in S. peasei; the verge (PV, fig. 43e) and muscle collar of S. avanaensis are larger than the equivalent structures in S. peasei (fig. 42d-e); and the pallial re- gion of S. avanaensis (fig. 43b) is slightly shorter, with a noticeably shorter and broader rectal kidney arm than in S. peasei (fig. 42c). The two species apparently have ecological separation, with S. peasei found almost without exception under rotting wood and stones, whereas, except at a very dry station, S. avanaensis lived in moss and lichens on tree trunks. More detailed collecting in Avana Valley on Rarotonga is needed to determine if they are actually sympatric at any point or if they are microallopatric in their distribution. In the Lau Archipelago of Fiji, for example, both S. inermis (Mousson, 1870) and S. adposita (Mousson, 1870) have been collected on Lakemba, Aiwa, and Yangasa Levu, and fragmentary material adequate for dissection of both species from a single island was available. The penis length of S. inermis is 0.8-1.0 mm. compared with 1.65 mm. for S. adposita, and the shells are very distinctive. The differences are larger than in the Rarotongan example. Sinployea adposita has a normal vergic papilla and a very small stimula- tory pad, whereas S. inermis has a large stimulatory pad and a small vergic papilla. On Mt. Lamlan, Guam (Station 137), both Himer- oconcha lamlanensis and H. fusca (Quadras & Mollen- dorff, 1894) were taken alive. The former has a short- ened penis (2.0 mm.) with major pilasters forming halfway from epiphallic pore to atrium, and the latter has a longer (2.7 mm.) penis with the major pilasters starting just below the epiphallic pore (compare fig. 103a-b, d, f). These differences would serve quite effec- tively as species recognition devices. Fortunately, differences between species in the Pacific Island Charopidae appear to be rather striking and uniform, although often difficult to characterize verbally because of extensive local variation. Basically the same procedures and philosophical approach were used to study both the Charopidae and Endodontidae, so that the results are directly compar- able in terms of species discrimination, classification, and phylogenetic interpretations. The greater generic diversity and usually lower specific diversity in the Pacific Island Charopidae are interpreted as resulting from multiple invasions from other areas at a more recent time than for the Endodontidae, plus the proba- ble fact that the Helicarionidae and Euconulidae may have been invading only slightly later in time than the Charopidae. Conceivably, this often arboreal group of Limacacea could have denied the opportunity for ex- tensive arboreal radiation to the Charopidae. PATTERNS OF MORPHOLOGICAL VARIATION Initial studies from 1961 through 1969 on both shell and anatomy were carried out at 100 x magnifica- tion or less using a Leitz dissecting microscope. Radu- lar mounts in Euparol were examined unstained using a Leitz Ortholux compound microscope under bright- field, dark-field, and phase-contrast illuminations. Subsequently, SEM observations on both shell and radular features resulted in a series of reports (Solem, 1970b, 1972a-c, 1973a-b, d, 1974, 1975, 1976a, 1977a-b) whose results are partially incorporated here. None of the available anatomical material was suitable for histological investigation. In many cases the apical genitalia of the BPBM material could not be studied because of the preservation technique utilized (Solem, 1976b, p. 19). The following survey of shell and anatomical vari- ation is less detailed than that presented for the En- dodontidae (Solem, 1976b, pp. 19-99). Instead of one phylogenetic lineage being represented, the Pacific Is- land Charopidae include several independent lineages. Instead of a monophyletic origin for the apertural bar- riers, several independent origins are involved. Thus, certain of the questions asked in the endodontid review are not relevant and were omitted. In terms of shell structure, the Charopidae on the Pacific Islands simply are more conservative. SIZE AND SHAPE VARIATIONS At least one specimen was seen of all the species treated in this monograph, so that the discussion that follows includes all the named taxa. Specimens that were indicated as probably representing new taxa, but which were not formally named, have been omitted from the comments and statistical analyses. No at- tempt at use of factor or multivariate analysis has been attempted because only the relatively crude standard shell measurements were available. The overall pat- tern of mean measurements of each species follows that of the Endodontidae (Solem, 1976b, pp. 20-21, figs. 6-10) except that they are more centrally clustered. Table I summarizes the basic distribution for nine var- iables or ratios. Although the extremes are well re- moved from the median, the distance between the first and third quartiles is relatively restricted. For shell height, diameter, and H/D ratio the first to third quar- tile range is 18%- 19% of the total range, for the whorl count it is only 13%, for apical cords only 16%, and for the number of ribs on the body whorl (assuming the maximum as 200) and body width it is 22.5%. Consid- TABLE I. - RANGE OF VARIATION IN THE PACIFIC ISLAND CHAROPIDAE. 1st Minimum Quartile 3rd Quartile Means of: Height 0. 48 1. 23 1. 51 1 .81 3, 69 Diameter 1. 07 2. 27 2. 76 3 ,49 7, 52 H/D Ratio 0. 365 0. 486 0. 523 .566 0. 801 Whorls 3 1/8+ 3 7/8 4 1/8- 4 3/8- 6 5/8 D/U Ratio 2. 03 3. 44 3. 94 4. 73 CLOSED Ribs on body whorl 18 .9 74 .0 91 .1 115.0 REDUCED Ribs /mm. Body Whorl Width 1.70 0.32 Apical Cords 6.5 7.66 0.76 9.5 11.25 0.90 10.5 14.61 1.15 12.0 REDUCED 2.06 22.0 ering that four phyletic lines are clumped, this is a quite narrow basic range of variation. It is also much less than in the Endodontidae (Solem, unpublished data). The extent to which this results from distortions of the brood chamber formation as opposed to the much larger number of taxa and greater age of the Endodont- idae is unknown. Body whorl contour in the Charopidae shows the same type of variation found in the Endodontidae but differs in frequencies. Table II contrasts the two families, using the same definitions as for the char- acter states in the Endodontidae except for omitting the brood chamber condition that has no Charopidae equivalent. Lateral compression, angulation, or cari- nation of the periphery is far more prevalent in the TABLE II. - BODY WHORL CONTOUR IN ENDODONTIDAE AND CHAROPIDAE. Total Number of Taxa: Endodontidae 38 43 21 Body Whorl Contour Laterally compressed Evenly rounded Flattened above and/or below rounded periphery Angled periphery 14 Carinated periphery 28 State not recorded for several taxa. Charopidae 6 41 39 1 3 PATTERNS OF MORPHOLOGICAL VARIATION Endodontidae. The Charopidae almost all have either an evenly rounded periphery or a mild to prominent, Sinployea canalis (Garrett, 1872) (fig. 49b), lateral flat- tening above a rounded periphery. A very few species, such as Sinployea planospira (Garrett, 1881) (fig. 46e), have a clearly laterally compressed periphery. Only Sinployea angularis (fig. 64e) has an angled periphery, whereas Himeroconcha quadrasi (Mollendorff, 1894) (fig. 105b), H. fusca (Quadras & Mollendorff, 1894) (fig. 105e), and especially Maafu thaumasius (fig. 76b) have carinated peripheries. These differences in peripheral contour obviously are not exclusive and have no value in trying to characterize the families, except to point out another area in which their average pattern of growth diverges. In the Endodontidae there was a noticeable effect on shell diameter and H/D ratio with changes in body whorl contour (Solem, 1976b, p. 23, fig. 14). The same pattern applies to the few Charopidae with drastically altered contours, but the numbers are too few to warrant statistical treatment. Because there are significant differences between the families in terms of spire protrusion, a comparison is given on pp. 38-40. The only really high-spired charopid is Ba humbugi (fig. 74b). The only truly sunken spire in the Charopidae is found in Roimontis tolotomensis (fig. 91a), three additional species have the spire depressed below the body whorl (Maafu thaumasius, fig. 76a-b; Himeroconcha lamlanensis, fig. 104d-e; Kubaryiellus kubaryi, fig. 89a-b), and 11 taxa have the spire barely emergent to slightly de- pressed or actually variable in the case of Lagivala demani (Tapparone-Canefri, 1883). The differences in spire elevation are so minor that no analyses of possi- ble effects on basic measurements are presented. Umbilical shape in the Charopidae is monoton- ously V-shaped, becoming saucer-shaped when more widely open as in Discocharopa (fig. 36c) and nearly U-shaped in the few species where it is narrowed (figs. 41c, 51f). There is none of the rich variation seen in the Endodontidae (Solem, 1976b, p. 27, fig. 16), hence only a few comments were presented in the family compari- sons (p. 40). Many of the Charopidae tend to develop a columel- lar and basal callus upon cessation of growth. This has the effect of covering over threadlike apertural bar- riers on the columellar wall and basal lip. Much of the variation in presence or absence of a threadlike col- umellar barrier recorded in Table V probably is di- rectly caused by the degree of basal callus formation. It shows particularly clearly in the illustrations of Pal- line biakensis (fig. 96b), P. micramyla (fig. 95d), P. no- terapalauana (fig. 94e), and Vatusila kondoi (fig. 82b). In complete contrast, taxa such as Kubaryiellus kubaryi (Mollendorff, 1900) and Russatus nigrescens (Mollendorff, 1900) (fig. 89b, e) totally lack any trace of such a callus. A final comment is required concerning the de- velopment of sulci on the whorls and the very unusual canalization of the sutural area. In the Endodontidae (Solem, 1976b, pp. 51-52) development of a subsutural and/or subperipheral sulcus is primarily associated with keel formation. In the two sharply keeled Char- opidae, Maafu thaumasius (fig. 76a-c) and Himero- concha quadrasi (Mollendorff, 1894) (fig. 105b), quite prominent sulci are present both above and below the periphery. In H. fusca (Quadras & Mollendorff, 1894) (fig. 105e), which is much less sharply keeled, only a weak supraperipheral sulcus is present. In the one species with angulated periphery, Sinployea angularis (fig. 64e), there is a clearly defined subsutural sulcus, and in both Vatusila kondoi (fig. 82b) and Sinployea rudis (fig. 47e), which have rounded peripheries, there are distinct supraperipheral sulci. They are the only equivalent to the unusual sulcus formation reported in the endodontid genus Anceyodonta (Solem, 1976b, p. 51, figs. 81d-e, 83, 86a, c, 88a, c, 89a). The functional significance, if any, of this last development is un- known. A most unusual modification involves canalization of the suture. This is more clearly seen in Sinployea proxima (Garrett, 1872) (fig. 46a-b) where it is accom- plished by slight detachment of at least part of the body whorl and creation of a narrow channel at the suture. This also occurs in Lagivala macroglyphis (Rensch, 1937) (fig. 81d-e), where the channel is formed by de- tachment to a greater angle and thus is shallower and wider, and to a barely detectable extent in L. micro- glyphis (Rensch, 1937). The same phenomenon is seen in the New Caledonian species Andre francia alveolus (Gassies, 1881) (Solem, 1961, pp. 454-456, fig. 12) and to a lesser extent in some Australian and New Zealand taxa. There is no equivalent change in the Endodont- idae. SHELL SCULPTURE Data on the basic method of sculpture formation, apical sculpture type, and sculpture spacing is pre- sented on pp. 40-41 as part of the family comparison discussion. In general the postapical sculpture in the Charopidae is relatively uniform in composition. It consists of major ribs, a few microradials in between, and fine microspiral segments that act as lateral sup- ports to the microradials (figs. Id-e, 2b, d), much in the same way as the buttresses to the microradials found in Ptychodon microundulata (Suter, 1890) (Solem, 1970b, pi. 59, figs. 9-11). The latter are quite different from the microspirals found in the Endodontidae, but this difference can be investigated only with the SEM. In some of the charopid species with reduced radial sculpture, such as Himeroconcha, these microspirals become enlarged into essentially continuous spiral cords. Again, this is far less specialized than in some of the situations found in the Endodontidae. As an example of the specializations in microspiral elements, compare Figures Id-e, 2d, and 3b, e. In Sinployea modicella (Ferussac, 1840) (fig. Id-e) the microspirals are on both sides of the radial riblets and tend to be joined medially and form minor cordlets. In FIG. 1. Shell microsculpture in Sinployea modicella (Ferus- sac). Station 893, Faatoai Valley, Moorea, Society Islands. BPBM 150377: a-b, apical and early postnuclear sculpture, note radial element on apex in a (295x) and secondary spiral "cording" in b (310x); c, detail of apical sculpture, note minute radial undula- tions and even nature of spiral cords (970 x); d, postnuclear sculpture between two major ribs on body whorl (960 x); e, detail of microsculpture, note uneven nature of microspiral buttresses (2,950x). PATTERNS OF MORPHOLOGICAL VARIATION 11 contrast, Tuimalila pilsbryi (fig. 2d) shows some of the same kind, but in the same inter-rib area it can have the microspirals as buttresses on only one side of the microradial. Palline biakensis (fig. 3b, e) and Micro- charopa mimula (fig. 4a) have very minor actual surface microspiral elements (fig. 3e), simply a very narrow support ridge leading up to the top of the mi- croradial whose upper edge then expands into an elon- gated bead (fig. 3e) upward along the riblet. These "beads" appear, when viewed at lower magnification (fig. 3b), to form microspiral "cords" that at optical magnification are equivalent in aspect to the serrated edges on the microriblets in the endodontid Austral- donta raivavaeana Solem (1976b, p. 33, fig. 23a-d). In both situations the spiral sculpture is barely visible at 100 x magnification, and determination of the compo- FIG. 2. Shell microsculpture in Sinployea peasei and Tuimalila pilsbryi: a-b, Sinployea peasei. Station 14, Avana Valley, Rarotonga, Cook Islands. FMNH 144613. a, apex and early postnuclear sculpture (320 x); b, suture between apex and first postnuclear whorl, note beaded secondary spiral "cording" and waved effect of apical cords caused by conforming to minute radial undulations on surface ( 1,065 x); c-d, Tuimalila pilsbryi. Station T-22, east side of Eua, Tonga. FMNH 152378. c, Apical and early postnuclear sculpture, note definite, widely spaced large radial swellings on apex, reduced size of spiral cords (320x); d, postnuclear sculpture on body whorl, note varying height of microspiral buttresses (l,065x). FIG. 3. Shell microsculpture in Palline biakensis. Hospitaal- grot, Biak, West Irian. Holotype. Rijksmuseum van Natuurlijke Historie, Leiden: a, apical and early postnuclear sculpture (275 x); b, 1st and 2nd postnuclear whorls (280x); c, detail of apical- 1st postnuclear whorl suture (930x); d, end of apical shell sculpture (920x); e, detail of sutural area between 1st and 2nd postnuclear whorls (910 x). 12 FIG. 4. Shell microsculpture in Microcharopa mimula. Cen- tral peak, Mothe, Lau Archipelago, Fiji. BPBM 78585: a, apex and part of 1st postnuclear whorl (295x); b, suture between apex and 1st postnuclear whorl (970x); c, center of apex (980x); d, center of apex (2,935x); e, detail of apical ridge (9,850x). 13 14 SOLEM: ENDODONTOID LAND SNAILS nents creating the visual effect requires l,000x SEM viewing. The transition between apical and postapical sculpture is quite abrupt (fig. 3d). There is a very short distance of stuttering apical growth, then a start of the typical postapical sculpture immediately. No dif- ferences in this pattern were seen in other species examined with the SEM. Rib counts were made on the body whorl of all unworn adult specimens. The pattern of mean rib counts and rib spacing is presented in Table I and also discussed in the family comparison section. The ques- tion of the utility of rib counts is covered by Solem (1976b, pp. 39-44), and justification of this is not re- peated here. Data on shell diameter and rib spacing correlations are presented in Table III. The correlation between shell diameter and rib spacing is less tight than in the Endodontidae (Solem, 1976b, p. 44, table XV), probably reflecting in part the smaller number of species involved. Rib reduction is discussed elsewhere (p. 9) and also is not as tightly correlated with size as in the Endodontidae (Solem, 1976b, pp. 47-50). In the Endodontidae and the Charopidae there are a few situations in which the major ribs have become grossly enlarged and widely spaced. Although none of the charopids equals the spectacular "wings" of Zyz- zyxdonta alata Solem (1976b, p. 466, fig. 198a-c), one does come close to the huge ribs of Cookeconcha stell- ulus (Gould, 1844) (Solem, 1976b, p. 218, fig. 93). Palikirus cosmetus (fig. 91d-f) from Ponape is the one example of spectacular rib enlargement in the Charopidae. Even when a keel is formed and the shell enlarged as in Maafu thaumasius (fig. 76a-c), the ribs themselves remain relatively small. When there is great size enlargement of the shell in such endodontid genera as Gambiodonta (Solem, 1976b, pp. 435-442, figs. 186-189), sometimes (G. mirabilis Solem, 1976; G. grandis Solem, 1976) there is clear rib enlargement. Variation in the number of apical cords between charopid species is relatively small and generally is TABLE III. - SHELL DIAMETER AND RIB SPACING IN THE CHAROPIDAE. Ribs/nun. LESS THAN 2 2.0-2.99 3.0-3.99 4.0-4.99 5.0-5.99 6.0-6.99 7.0-7.99 8.0-9.99 10.0-12.99 13.0-19.99 MORE THAN 20 Number of Taxa 2 1 2 5 4 6 5 14 26 14 6 Mean and Range of Shell Diameter in mm. 6.06(4 2.40 4.00(3 4.88(3 3.29(2 3.45(2 2.69(1 3.06(1 2.61(1 2.29(1 1.91(1 59-7.52) 29-4.70) 49-5.68) 11-4.30) 17-5.23) 75-3.49) 65-4.63) 73-3.74) 40-2.95) 07-3.21) greatly exceeded by the variation within species. A range of four to six within a population or species is not unusual where the mean is 11 or 12, dropping to a range of three to four where the mean is eight or nine cords. Such variation in part reflects accidents of post- nuclear whorl attachments that can change the count by one or two cords, which is partly what happens in terms of cord fragmentation near the upper suture and is partly actual variation in cord width and spacing. Because the intrapopulational variation is so large, lit- tle importance is attached to minor mean differences. Data on 83 of the 95 taxa were available. Of the re- maining 12, Discocharopa has only radial apical sculpture, three had an apical sculpture too fine to count optically, five had the apex too worn, and three had the apical whorls missing or were worn fossils. Table I presents the dispersal of the species means, which shows a rather tight central cluster. Of the five species with a mean of less than 8.5 spiral cords, three (Roimontis tolotomensis, Himeroconcha lamlanensis, and Kubaryiellus kubaryi) have a moderately de- pressed to sunken spire, Jokajdon callizonus (Mollen- dorff, 1900) (fig. 92a-c) may be involved in secondary size reduction, and only Vatusila vaitupuensis (fig. 85a-c) shows no special features explaining the reduc- tion in number of apical cords. At the other extreme, Lauopa mbalavuana (fig. 77a-c) with 22 cords, Mi- crocharopa mimula (fig. 4a) with a mean of 17.8, Sinployea irregularis (Garrett, 1872) (fig. 65d-f) with 18.5, S. recursa (fig. 64a-c) with 15.9, S. angularis (fig. 64d-f) with 14.6, and Palline biakensis (fig. 96a-c) with 14.0 cords show no clear indication of why the numbers have increased. Even the very elevated spire of Ba humbugi (fig. 74b) did not make a noticeable difference in apical cords (mean 12.0). The number is undoubtedly much higher in the species of Tuimalila (fig. 2c), but these cords are nearly suboptical in size and are partly obscured by the radial swellings. Variation in the height of the apical cords can be studied easily with the SEM. A few examples are shown in Figures 1-3. The situation in Sinployea modicella (Ferussac, 1840) is typical (fig. la-c), with the height of each cord (mean 9.2 cords) being slightly less than its width. There is an obscure radial rugosity as a secondary feature. Within the same genus, Sinployea peasei (fig. 2a-b) (mean 11.4 cords) shows a reduction in both height and width of each cord, al- though the underlying rugosity is just as clearly out- lined. In Tuimalila pilsbryi (fig. 2c) the spiral cords are about the same size as in Sinployea peasei, but the secondary radial swellings confuse the viewer using optical equipment. Finally, in Palline biakensis (fig. 3a-d) the spiral cords are much higher and narrower, and there is no trace of the radial rugosities seen in the Sinployea. Only two Pacific Island charopids show a different apical sculpture. Microcharopa mimula (fig. 4a-e) has spiral apical ribbing, but this is broken up into a series of short segments with peculiar open ends (fig. 4d-e). PATTERNS OF MORPHOLOGICAL VARIATION 15 At optical magnifications the sculpture appears to be simple spiral cords, but obviously it is much more com- plex. The same sculpture is found in the North Ameri- can Radiodiscus millecostatus (Pilsbry & Ferriss, 1906) (Solem, 1977b, p. 152, figs. 7-8) and an unnamed Western Australian genus (Solem, unpublished data). These genera are grouped in the charopid subfamily Rotadiscinae H. B. Baker (1927, p. 228). Discocharopa aperta (Mollendorff, 1888) (fig. 5a-c) has an apical sculpture of fine radial ribs that become very crowded (fig. 5a) near the end of the apex. Higher on the apex (fig. 5b-c) there are irregularly spaced periostracal spiral folds that add a partial spiral element. The post- apical whorls (fig. 5d-f) show no trace of any spiral sculpture. In summary, the Pacific Island Charopidae mostly have a very characteristic and conservative pattern of sculpture. The apex has a variable number of spiral cords to which radial swellings are occasionally added. The postapical whorls have major radials whose spac- ing and number are somewhat correlated with shell diameter, finer microradials in between, with very fine microspiral units that buttress or connect the microra- dials. Most of this sculpture is formed by the perio- stracum (p. 137). Reduction in prominence or loss of sculptural features is an apparently rare event. The situation is much more complex in relation to the New Zealand and Australian charopid taxa. APERTURAL BARRIERS Traditionally, those endodontoid taxa with aper- tural barriers were placed in the form genus En- dodonta, and those without barriers, in the form genus Champa (Solem, 1976b, pp. 118-119). The discovery that taxa with apertural barriers showed family-level differences in anatomy is one of the more significant results of this study. Solem ( 1973b) reviewed the bar- rier frequency and structure in the two families and concluded that although the endodontoid barriers are of common origin, those found in the Pacific Island Charopidae have evolved at least four times. Thus, in- stead of considering monophyletic variation, as in the equivalent discussion of the endodontid barriers (Solem, 1976b, pp. 52-72), an obviously polyphyletic situation exists. Such barriers are a common phenomenon in many land snail families (Solem, 1972c), but the timing of their appearance in ontogeny, methods of formation, state in adulthood of the individual, positions in the aperture, length, and numbers differ radically. They are analogous rather than homologous developments. Retention of specialized terms for the different structures found in separate families is fully justified. I have chosen to use a uniform terminology for the bar- riers in the Endodontidae and Charopidae, despite their independent origins, for ease in comparisons and because the barriers have the same patterns of change and growth. In both families the barriers appear at or very shortly after hatching, are added to anteriorly and resorbed posteriorly during subsequent growth, and often effectively narrow the aperture in adulthood. Recognition of parietal, columellar, and palatal zones in the aperture, the differences between major barriers and small traces, and the numbering sequence [parietals (upper to lower), columellars (upper to lower), palatals (lower to upper), all in numerical se- quence regardless of possible homology] have been given by Solem (1976b, pp. 52-54, figs. 36-38). The occurrence of barriers in the Charopidae is neither taxonomically nor geographically random: 32 (33.7%) of the 95 taxa have parietal and 27 (28.4%) have palatal barriers. 1 These are clustered in 11 of the 20 genera. Lauopa, Discocharopa, Ladronellum, and Roimontis are monotypic, Jokajdon and Palikirus have two species, Graeffedon and Palline have three, Sem- perdon has five, and Lagivala and Vatusila have six species. No species in these genera are without bar- riers, except for Palikirus ponapicus (Mollendorff, 1900), which is known from a single, fungus-covered specimen and was associated with Palikirus as a con- venience, and great secondary reduction of the barriers in Vatusila niueana (fig. 83e). Five of the 32 species are basically extralimital: Lagivala macroglyphis (Rensch, 1937) and L. microglyphis (Rensch, 1937) from New Britain; L. demani (Tapparone-Canefri, 1883) from Ambon, Timor, and parts of West Irian; Palline biaken- sis from West Irian; and Discocharopa aperta (Moll- endorff, 1888), which ranges from the Philippines to Society Islands. The three Graeffedon are limited to Western Samoa; the Lau Archipelago has two species (Vatusila kondoi and V. nayauana) on Nayau and Lauopa mbalavuana on Vanua Mbalavu and two species on Viti Levu (Lagivala vivus and L. minus- culus). Vatusila tongensis on Eua, Tonga; Lagivala davidi (Ladd, 1968) fossil on Funafuti and Vatusila vaitupuensis live on Vaitupu, Ellice Islands; V. niueana on isolated Niue Island; and V. eniwetokensis (Ladd, 1958) fossil on Eniwetok in the Marshall Islands complete the western distribution. A second cluster of taxa occurs on some of the Caroline, Palau, and Mariana Islands. On Ponape there are Palline mi- cramyla, Jokajdon callizonus (Mollendorff, 1900), J. tumidulus (Mollendorff, 1900), Palikirus cosmetus (Mollendorff, 1900), and Roimontis tolotomensis. On the islands of Palau, the three subspecies of Palline notera, Semperdon uncatus, S. xyleborus, and S. kororensis (Beddome, 1889), are found. In the Mariana Islands S. heptaptychius (Quadras & Mollendorff, 1894) and Ladronellum mariannarum (Quadras & Mollendorff, 1894) live on Guam, and Semperdon rotanus occurs on Rota and the northern tip of Guam. 'This differs from the figure cited by Solem (1973b, p. 301) be- cause Discocharopa aperta (Mollendorff, 1888) has some populations with and many without a parietal barrier, so this species was omitted from that calculation. FIG. 5. Details of shell sculpture in Discocharopa aperta (Mollendorff). Busuanga, Philippine Islands. FMNH 18725: a, shell sculpture at apical-postnuclear boundary (l,000x); b, apical and postapical sculpture (l,035x); c, details of apical sculpture (3,000x); d, postnuclear sculpture (l,015x); e, postnuclear sculpture showing serrated edging (3,000x); f, detail of postnuclear sculpture showing absence of spiral elements (9,950x). 16 PATTERNS OF MORPHOLOGICAL VARIATION 17 Except for the scattered occurrence of Dis- cocharopa aperta (Mollendorff, 1888) in the Austral Is- lands and on Borabora, Society Islands, the entire Cook and Society Islands area of great charopid diversity is free from barrier-bearing genera. Only the one taxon is known from Tonga, although several Sinployea and Tuimalila live in this group. Both the Lau Archipelago and Viti Levu have more taxa without than with bar- riers, as does Samoa. In Micronesia, Guam has the en- demic barrierless Himeroconcha, and Ponape has two genera, Kubaryiellus andRussatus, that lack barriers. The Palau group is unique in having only barrier- equipped taxa in both the Endodontidae and Charo- pidae. On some other Caroline Islands, there are barrier-free taxa, and on both Funafuti and Vaitupu there are species of Sinployea. The extension of barrier-equipped taxa to the Marshall Islands repre- sents their only penetration beyond the barrier-free taxa. A summary of the evidence that the barriers in these genera represent independent derivations was given in Solem (1973b). The uniform pattern of micro- armature and growth in the Endodontidae was re- viewed both in Solem (1973b) and in more detail sub- sequently (Solem, 1976b, pp. 54-66, figs. 39-42). In brief, the tops of the endodontid barriers normally are modestly expanded into a cordlike shape, with very fine triangular points (Solem, 1976b, figs. 39e, 40, 41a-c, 42a-f) that are added to the surface. They are not parts of basic crystal layers that simply extend outward, but are an extra feature. Checking the re- sorption edge of a barrier (Solem, 1976b, p. 64, fig. 40) shows that they, at times, will be covered up by sub- sequent growth as the height of the barrier is in- creased. The patterns in the Charopidae are only partly studied. At the time that most of the material used was on loan for study (1962-1967), the SEM was not yet available. In some of the genera, Lagiuala, Vatusila, and Graeffedon, specimens are few in number and the barriers too deeply recessed for adequate SEM study without breaking the shell. Thus the following com- ments must be viewed as only introductory. In Jokajdon callizonus (Mollendorff, 1900) (fig. 6a-d) from Ponape, views of the third palatal barrier demonstrate the basic structures in one species com- plex. In vertical view of the posterior descending face, crest, and upper anterior descending face (fig. 6a), the etched resorption plane is on the left, whereas the right shows the larger platelets on the upper anterior slope of the barrier. At higher magnification (fig. 6b), the mixture of large and small platelets on the upper an- terior face is seen in vertical view. In Figure 6d, the same area is shown in lateral view to demonstrate that the platelets are simply prolonged crystals from the growth face of the barrier. Figure 6c, at the point where the sharp anterior descension of the barrier ends, shows that there is an abrupt shift from large to small platelets. At this point, the combined major func- tion of gripping surface for extending the body from the shell aperture and providing setal catching projection against arthropod entry becomes much less significant, and hence the probable shift from large to small platelets. Looking at the first palatal barrier in the Western SamoanGraeffedongraeffei (Mousson, 1869) (fig. 7a-c) shows a second pattern of structure. The specimen was taken alive, and the shell was partly fractured during dissection. The barrier surface was not worn, so that the structure of long, straight, projecting ridges across the anterior descending face of the barrier is normal. There is distinct variation in the length of the ridges (fig 7b-c). The most intriguing aspect is seen on the side of the barrier (fig. 7a). The individual pockets hold platelets at a variety of angles, sometimes protruding, sometimes almost vertical in orientation. I consider it possible that they represent fragments from the ad- sorption surface that have been transported and re- fixed in a near anterior position. Unfortunately, specimens of Graeffedon are quite rare, and it is im- probable that material sufficient to test this idea will become available for study. The Palau Island Semperdon xyleborus (fig. 8a-e) has long parietal barriers, with the upper parietal posteriorly bifurcate (fig. 8a). The upper edge of the third parietal (fig. 8b-c) shows a few broad, projecting plates that point anteriorly. Some of them stretch across almost the entire edge. In lateral view (fig. 8d), the edge of the fourth parietal shows that the plates become much smaller on the side of the barrier and that the large plates are growing from the barrier sur- face as it slants slightly downward. They are partly elevated from the surface. The basic difference between the three types can be summarized as: ( 1 ) crystal layers simply continued directly forward and broken into microplatelets (Jo- kajdon); (2) anterior added ridges that may be in part transported as large crystals (Graeffedon); and (3) broad ridges angled up from the surface layers (Sem- perdon). Despite similar appearance and function, these are actually three different ways of reaching the same end result, i.e., a series of microprojections that provide gripping surfaces for the advancing mantle col- lar during emergence of the body from a fully with- drawn position. They are just as effective as the fourth solution, the triangular points in the Endodontidae, but are not as elegantly constructed. Coupled with the very obvious size and position differences of the bar- riers in the three charopid examples cited, which are typical of their anatomical groupings, I conclude that they are of independent origin. When compared with the apertural barrier microstructure in other families (Solem, 1972c), the charopid experiments show greater individual variation and are much less "finished" in appearance, even though the use of angled plates is a more common pattern than the addition of points or barbs on the surfaces. The presence of barriers in the Charopidae is a polyphyletic and secondary character, rather than the norm as in the Endodontidae. Thus, the comparative SOLEM: ENDODONTOID LAND SNAILS FIG. 6. Microstructure of largest palatal barrier mJokajdon callizonus (Mollendorff). Station 118, Ponape, Caroline Islands. BPBM 154161: a, vertical view onto top of largest palatal at l,030x, note weak etch marks on upper left and the clear cross-strata pattern on entire posterior (left margin); b, area of sharp anterior slope in vertical view at 3,200x; c, area at bottom of sharp anterior descension showing transition from large crystal plates to small crystals at 3,125x; d, lateral view at 3,050x of large crystal area showing how these plates are continuations of the vertical growth layers. treatment will be different than in Solem (1976b). Correlations among barrier condition and other shell features are not as simple and direct as in the En- dodontidae. For example, the relative degree of aper- ture narrowing in the Endodontidae (Solem, 1976b, p. 71, table XLVIII) and Charopidae (table IV) is quite different. The percentages of the barrier-possessing taxa that are intermediate in closure differ greatly. In the Endodontidae, 57.5% are intermediate, but only 25.0% in the Charopidae. The latter have the barriers TABLE IV. - DEGREE OF APERTURAL NARROWING BY BARRIERS Strong Moderate Weak Not Total Taxa Endodontidae 19.9% 40.9% 16.6% 22.6% 181 Charopidae 43.7% 9.4% 15.6% 31.3% 32 PATTERNS OF MORPHOLOGICAL VARIATION 19 FIG. 7. Surface of largest palatal barrier in Graeffedon graeffei (Mousson). Station 39, Mt. Solaua, Upolu, Western Samoa. FMNH 153420: a, angled view of front edge and inner side showing different lateral and anterior surface texture (305 x); b, anterior edge at 1,100 x showing projecting platelets on surface of descending edge; c, ver- tical view onto descending edge surface at 3,150x . either strongly constricting the aperture (figs. 98b, e, 99b, d, e), or they are essentially ridgelike remnants (figs. 85b, lOlb). In both families, more species have parietal bar- riers than either columellar and palatal, and no species has either of the latter without also having the former. Parietal barriers The number of parietal barriers ranges from the single barrier that may be present or absent in Dis- cocharopa, to the four large barriers plus three traces found in Semperdon xyleborus (fig. 98e). The only vari- ation in major barrier numbers occurs in Semperdon heptaptychius (Quadras & Mollendorff, 1894), which normally has three parietals and one accessory trace. Occasionally (5.1%), the trace and third parietal are missing, rarely (0.3%), a fourth parietal is developed, and rarely (3.3%), the trace is missing. In Semperdon xyleborus, 84.6% have three traces, 15.4% have two traces. Occasionally, a parietal trace will be present in Palline notera notera and P. n. gianda, and there is one superior trace present in S. rotanus (fig. 99d). Other- wise, the number of parietal barriers is constant. Barrier length partly correlates with height. The method of determining length is explained by Solem (1976b, p. 70, fig. 43). Where the parietal is a thread- like remnant, it is Vie to Vs whorl long (figs. 83e, 85b, 95d, lOlb), except in Palline biakensis (fig. 96b) and Roimontis tolotomensis (fig. 91b) where the length reaches 3 /ie of a whorl. In Vatusila nayauana (fig. 82e), Semperdon uncatus (fig. 97b), and Palikirus cosmetus (fig. 91e), the parietal is a raised ridge, but still short. The only taxa with high, short barriers are the races of Palline notera (figs. 94b, e, 95a) and Semperdon xyleborus (figs. 8a, 98e). The remaining barriers are V* to 3 /ie of a whorl long, except for several species of Lagivala where they extend to or beyond the line of vision from the aperture (L. microglyphis, fig. 81b; L. macroglyphis, fig. 81e; L. minusculus, fig. 80e). These species are very small in size, with mean diameters of 1.55-1. 73 mm. Structure of the barriers is highly varied. The threadlike and raised ridge taxa present no unusual features and have been mentioned above. Where there is a single high barrier, such as Lauopa mbalauuana (fig. 77b), Discocharopa aperta (Mollendorff, 1888) (fig. 35a-b), and Lagivala minusculus (fig. 80e-f), it is a high thin blade, usually with abrupt anterior descen- sion. When there are two or three high lamellae, the typical shape found in most of the Endodontidae is usual: an abrupt posterior resorption face, elevated middle with slightly to strongly expanded upper edge, and then gradual anterior descension. The second and fourth parietals in Semperdon xyleborus (fig. 8a) are typical. Several taxa have a modification of the first parietal that also was found in some Endodontidae. It can become deflected downward on the posterior por- tion in Jokajdon callizonus (Mollendorff, 1900) (fig. 92b), Semperdon heptaptychius (Quadras & Mollen- FIG. 8. Structure of apertural barriers in Semperdon xyleborus. Station 184, Peleliu. Palau Islands: a, parietal barriers at 120x, note posterior bifidity of upper (1st) parietal: b, upper edge of 3rd parietal at 950x; c, upper edge of 3rd parietal at 2,850x; d, upper edge of 4th parietal at 2,910x; e, lower palatal barriers at 326x . Anterior side at left margin in each photograph. 20 PATTERNS OF MORPHOLOGICAL VARIATION 21 dorff, 1894) (fig. 99b), and S. rota nut* (fig. 99d); it is bifid in Vatusila tongensis (fig. 83b) or posteriorly bifurcated in V. kondoi (fig. 82b), Jokajdon tumidulus (Mollendorff, 1900) (fig. 92e), Semperdon kororensis (Beddome, 1889) (fig. 98b), and S. xy/ebon/s (fig. 8a, upper barrier). Similar bifurcation was found also in three lineages of the Endodontidae several Cooke- concha sand Endodonta (Solem, 1976b, p. 379, fig. 167a, g), some Min idon ta (Solem, 1976b, p. 138, fig. 65b,d-e, g; p. 149, fig. 71c), and some Anceyodonta (see Solem, 1976b, p. 57). The reason for these repetitive bifurca- tions, usually of the upper parietal barrier, is un- known. The basic similarity in shape between the larger parietals of the Charopidae and Endodontidae proba- bly indicates that this is an efficient shape, not only in terms of continual anterior addition and posterior re- sorption, but also in functioning during body extension of the animal. None of the Charopidae show the splitting into threadlike traces of the barriers that was common in several lines of the Endodontidae (Solem, 1976b, p. 57; p. 332, fig. 144a-c). The Charopidae show instead a quite conservative shape and number pattern of the parietal barriers. Columellar barriers Of the 32 taxa with parietal barriers, 13 lack any columellar barrier. This part of the shell is missing in Vatusila eniwetokensis (Ladd, 1958), so that its situa- tion is unknown. Semperdon xyleborus and S. kororen- sis (Beddome, 1889) (fig. 98b-c, e) have two prominent columellar barriers, and the former species normally has two accessory traces above the upper columellar. Six taxa, Lagivala macroglyphis (Rensch, 1937) (fig. 81e-f), L. microglyphis (Rensch, 1937) (fig. 81b), Jokajdon callizonus (Mollendorff, 1900) (fig. 92b), J. tumidulus (Mollendorff, 1900) (fig. 92e-f), Graeffedon savaiiensis (fig. 88b), and Pa/line notera notera (fig. 94b) have only one columellar. This varies from a threadlike ridge in the last two, to high and sometimes sharply ascending barriers in the other four species. The remaining 10 taxa (table V) show intraspecific variation in the number or absence of columellar bar- riers. When the variation is present or absent, in all but Lagivala demani (Tapparone-Canefri, 1883), the barrier is minute and threadlike. Its absence would mostly result from being covered by the basal callus. In Lagivala demani (Tapparone-Canefri, 1883) the varia- tion is geographic, with a large columellar present in specimens from Timor, medium-sized in Ambon and Aru Islands shells, and very small to absent in Biak and West Irian examples. In Semperdon rotanus and S. heptaptychius (Quadras & Mollendorff, 1894) (fig. 99e) the second columellar, when present, is a small thread- like trace above the larger columellar that is a high ridge parallel to the plane of coiling. A large form of the latter species (fig. 99b) occurs in which the columel- lar barrier is absent. TABLK V. - PERCKNTACK DISTRIBUTION OF COLUMKI.I.AR BARRIER NUMBERS. Vatusila kondoi V. nayauana Lagivala demani Graeffedon graeffei Palline notera gianda P. notera palauana Number of Columellar Barriers 1 5.6 94.4 40.0 60.0 * * 30.0 70.0 84.0 16.0 95.0 5.0 83.0 17.0 46.5 0.3 42.1 3.5 96.5 53.5 56.4 Semperdon uncatus S . rotanus ^. heptaptychius Ladronellum roar iannarum * Geographically variable. In most taxa the columellar barrier lies parallel to the plane of coiling. Semperdon xylcborus and S. kororensis (Beddome, 1889) have one barrier slanting downward (fig. 98b, e); Jokajdon tumidulus (Mollen- dorff, 1900) (fig. 92e), Lagivala microglyphis ( Rensch, 1937), and L. macroglyphis (Rensch, 1937) (fig. 81b, e-f) have it twisted up toward or almost onto the parietal wall; and in others, such as Jokajdon cal- lizonus (Mollendorff, 1900) (fig. 92b), it is a deeply re- cessed, triangular knob. Thus, the 18 species of Charopidae known to have a columellar barrier show almost the same total range of variation reported for the Endodontidae (Solem, 1976b, pp. 57-59). Palatal barriers Five of the 32 taxa with parietal barriers lack palatals Lauopa mba/auuana (fig. 77b),Discocharopa aperta (Mollendorff, 1888) (fig. 36b), Palline biakensis (fig. 96b), Roimontis tolotomensis (fig. 91b), and Palikirus cosmetus (fig. 91e). Except for the Roimontis. which has two threadlike parietals, all of them have only a single, short lamellar parietal. Because of the missing palatal wall in the holotype and only known example, the palatal barrier situation in Vatusila eniwetokensis (Ladd, 1958) is unknown. In Ladronel- lum mariannarum (Quadras & Mollendorff, 1894) (fig. lOlb), many Vatusila nayauana (fig. 82e), and V. niueana (fig. 83e) there is a modest ridge to barely detectable trace on the middle of the palatal wall. Semperdon uncatus (fig. 97b) normally has one low barrier on the basal lip, but in 17% of the individuals this is missing, and Palline micramyla (fig. 95d) has two low lamellar barriers on the basal lip. Lagivala minusculus (fig. 80e) has a rather unique ridge with a lateral buttress on the basal lip, and Vatusila tongensis (fig. 83b) has a very large basal lip barrier that is bifurcated or split into two parts plus a very unusual 22 SOLEM: ENDODONTOID LAND SNAILS upper raised callus that is 0.25-0.36 mm. wide. Vat- usila vaitupuensis (fig. 85b) has a midpalatal thread- like trace and a broader ridge on the basal lip very near the columellar margin. The remaining species have a more typical array of normally three to seven palatal barriers, sometimes with additional accessory traces present. The position- ing of these barriers varies from at the lip edge in Sem- perdon (figs. 98b-c, e, 99b, d, e),Palline notera and its subspecies (figs. 94b, e, 95a), Graeffedon (figs. 88b, e, 86b), and Lagivala microglyphis (Rensch, 1937) (fig. 81b), to partly recessed in Jokajdon callizonus (Moll- endorff, 1900) (fig. 92b), progressively recessed in J. tumidulus (Mollendorff, 1900) (fig. 92e), moderately recessed in Vatusila kondoi (fig. 82b) and Lagivala macroglyphis (Rensch, 1937) (fig. 81e), and deeply re- cessed in both L. vivus (fig. 80b) and L. davidi (Ladd, 1968) (Ladd, 1958, pi. 30, figs. 13-15). The shape of the barriers is rather variable among these species, although constant within a particular species. A simple crescent-shape with modestly to moderately expanded upper edge is found in Lagivala vivus (fig. 80b), L. macroglyphis (Rensch, 1937) (fig. 81e), andL. demani (Tapparone-Canefri, 1883); a more expanded upper edge, in L. microglyphis (Rensch, 1937) (fig. 81b) and Graeffedon pricei (fig. 88e); typical structure, in Vatusila kondoi (fig. 82b); very high, in Graeffedon graeffei (Mousson, 1869) (fig. 86b) and G. savaiiensis (fig. 88b); typical, in the races of Palline notera (figs. 94b, e, 95a); and bulbous-edged, again in Semperdon rotanus and S. heptaptychius (Quadras & Mollendorff, 1894) (fig. 99b, d-e). In Semperdon xyle- borus (figs. 8e, 98e), the crescent is anteriorly elon- gated, a process continued in the two species of Jokaj- don (fig. 92b, e,) and reaching the typical parietal bar- rier shape in Semperdon kororensis (Beddome, 1889) (fig. 98b). The predominance of the crescent-shaped palatal barriers outlined above is very different from the situa- tion in the Endodontidae. In that family (Solem, 1976b, p. 60, table XXXII) only five species had crescentic palatals, whereas 130 had bladelike barriers. In the Charopidae, only the two Jokajdon, Semperdon koro- rensis (Beddome, 1889), and possibly S. xyleborus could be considered to have a bladelike palatal barrier form. Intraspecific variation in the number of palatal barriers is summarized in Table VI. The great varia- tion found in the two Semperdon exceeds that found in any of the Endodontidae (Solem, 1976b, p. 67, table XL). None of the ratios in Table VI convincingly suggests genetic dominance. Assigning the most fre- quent percentage to represent that species, the pattern of palatal barrier numbers is summarized in Table VII. It is evident that there is considerable variation within genera, particularly Lagivala, Graeffedon, Palline, and Semperdon. Even within genera, there is no clear cor- relation between number of barriers and shell size. Only Palline notera shows a partial correlation, in that the two larger subspecies have fewer and smaller-sized TABLE VI. - PERCENTAGE DISTRIBUTION OF PALATAL BARRIER NUMBERS Taxon Number of Palatal Barriers 012345 Vatusila kondoi 11.1 88.9 V. nayauana 40.0 60.0 Lagivala demani * 20.0 80.0 10.0 75.0 10.0 5.0 16.0 68.0 16.0 1.2 3.0 19.8 55.0 1.2 80.0 13.1 3.9 1.5 92.0 21.0 0.3 8.0 Palline notera palauana V_. ri. gianda Semperdon uncatus 17.0 83.0 S. rotanus S. heptaptychius . xyleborus * Geographically variable. palatals than the nominate race, which averages 0.24-0.41 mm. less in diameter than the others. There is also no clear correlation between number of palatals and degree of apertural narrowing. The presence of palatal traces is sporadic. Several taxa will have one or two traces between lower pairs of palatal barriers: Lagivala microglyphis (Rensch, 1937), Graeffedon graeffei (Mousson, 1869), G. pricei, Palline notera gianda, Semperdon rotanus, and S. hepta- ptychius (Quadras & Mollendorff, 1894). Other taxa will have three traces above the upper palatal: Jokajdon callizonus (Mollendorff, 1900), Semperdon kororensis (Beddome, 1889), and most examples of S. xyleborus. A few of the latter will have four (7.5%) or five (7.5%) traces above. Jokajdon tumidulus (Mollendorff, 1900) normally has three traces above the upper palatal and three (rarely two) between lower pairs. The same pat- TABLE VII. - PALATAL BARRIER NUMBERS IN CONSTRICTED APERTURE CHAROPIDAE. 3 Palatals Vatusila kondoi Lagivala demani Jokajdon callizonus J_. tumidulus 4 Palatals Lagivala davidi Graeffedon pricei Palline notera gianda P_. _n. palauana Semperdon heptaptychius 5 Palatals Lagivala microglyphis U macroglyphis Graeffedon graeffei Palline notera notera 6 Palatals Lagivala vivus 7 Palatals Graeffedon savaiiensis Semperdon rotanus S. xyleborus S. kororensis PATTERNS OF MORPHOLOGICAL VARIATION 23 tern of traces above the upper palatal is seen in the Endodontidae (Solem, 1976b, p. 59) and was inter- preted as a space-preserving function, because the great size of the first parietal and the need to keep this space open for withdrawal and extension of the buccal mass and foot combine to make reduction in size of palatal projections in this area advantageous. As in the Endodontidae, the normal pattern of size change is for the palatal barriers to decrease in height as they ascend the palatal wall. An obvious exception will be reduced size of the first palatal in situations where it is right at the columellar margin (fig. 99d) or where there is large size to the columellar and/or lower parietal barriers (figs. 86b, 92b, e). So few taxa show either recession of the palatals or major upper edge expansion of the barriers that discussion of shell corre- lations separately is not attempted. Most of the com- ments made about general features of the palatal bar- riers in the Endodontidae (Solem, 1976b, pp. 59-63) seem to apply to the few Charopidae also. Despite only 27 charopids having palatal barriers, they show a wide range of variations that mostly paral- lel the situations found in the Endodontidae. The biggest difference lies in the basic shape of the bar- riers: bladelike in 86.1%, ridged in 10.6%, and crescent-shaped in 3.3% of those Endodontidae with palatal barriers compared with bladelike in 15.4% (four), a lamellar or unusual ridge in 23.1% (six), and crescent-shaped in 61.5% (16) of the 26 charopid taxa for which data on the palatal barrier structure were available. Judging from the size of the parietal barriers (fig. 85e), Vatusila eniwetokensis (Ladd, 1958) probably had large palatal barriers, but the entire outer wall of the only known specimen is missing, and thus the state of the barriers is unknown. Patterns of barrier variation Despite the lack of size and shape correlations with variation in the barriers found in the Charopidae, a few comments can be made concerning general pat- terns within a particular group. In Lagivala (figs. 80-81) the variation is geographic, with the Bismarck Archipelago species having the strongest apertural barriers, progressive reduction occurring in both L. demani (Tapparone-Canefri, 1883) to the west and in the Fijian and Funafuti taxa to the east (fig. 80b, e). Vatusila shows no size correlation, with the largest (V. tongensis, fig. 83b) and one of the smaller (V. eniwetokensis, fig. 85e) having quite large barriers, and both large (V. vaitupuensis, V. niueana, figs. 83e, 85b) and small (V. nayauana, fig. 82e) having reduced barriers. There does appear to be some geographic con- cordance, with Tonga and Lau species having larger barriers than taxa from the more isolated Vaitupu and Niue localities. The fossil record of a large-barriered form on Eniwetok in the Upper Miocene gives an his- torical perspective to the group that otherwise is lack- ing in the Charopidae. The reasons for the reduction in barrier size in the races of Palline notera on Koror (palauana, fig. 94e) and Babelthuap (gianda, fig. 95a) compared with the nominate race on Peleliu (fig. 94b) is unknown, al- though slight size differences are involved. Barrier size variation in Semperdon shows no correlations that I could detect. The unconstricted aperture of the mean diameter 2.83 mm. S. uncatus (fig. 97b) contrasts with the highly constricted aperture in the mean diameter 2.88 mm. S. xyleborus (fig. 98e) and mean diameter 4.59 mm. S. kororensis (Beddome, 1889) (fig. 98b). All of these species have been dissected and all occur on Koror. The differences in barrier prominence cannot be explained at present. Summary of barrier variations Despite the different microstructure on the barrier edges and the evident polyphyletic origin of the aper- tural barriers in the Charopidae, in general they paral- lel those of the Endodontidae. The retention of parietal barriers extending to the lip edge, general shape, and multiple bifurcations of the parietals all suggest paral- lel function. Given the similar habitat and overlapping shell size, plus the same basic growth structure, this is not surprising. The difference in shape frequency of the palatal barriers, a greater tendency for loss or reduc- tion in the barriers, and the general shorter length of the charopid barriers are real average differences but cannot be used as criteria for family-level separation because there is so much overlap between families. SUMMARY OF SHELL VARIATIONS With few exceptions the range of variation in the shells of the Charopidae is less than that found in the Endodontidae. Major differences between the two fam- ilies are reviewed on pp. 40-44. Here it is sufficient to point out that the formation of sculpture by the periostracum rather than the calcareous layers, lack of umbilical and growth modifications to form a brood chamber, and secondary derivation of apertural bar- riers in the Charopidae have had profound effects on the range of shell variation. GROSS ANATOMY Including the Punctum sp. from Tahiti, Sinployea sp. from Borabora, and Sinployea sp. from Saipan, there are 98 species-level taxa considered in this monograph. Of these, 51 (52.0%>) belong to one genus, Sinployea. It was possible to see at least fragmentary soft parts for 43 (43.9%) of the total taxa and 21 (41.1%) of the Sinployea. The taxonomic distribution of the dis- sected material is given in Table VIII. Of the 21 gen- era, six genera, five of them monotypic, could not be dissected. Only Lagivala, with six species, represents an undissected speciose group. The taxonomic coverage of dissected material is broader than that given the Endodontidae (Solem, 1976b, p. 73, table L), where only 58 (32.4%) of the 179 taxa seen were dissected. This reflects both a lesser degree of extinction for the 24 SOLEM: ENDODONTOID LAND SNAILS TABLE VIII. - PHYLETIC REPRESENTATION OF DISSECTED TAXA. Total Taxa Number Dissected Family Punctidae Punctum Family Charopidae 1 1 51 1 1 1 2 6 6 3 1 1 1 1 2 2 5 5 1 A 98 21 1 1 1 1 1 1 1 2 2 5 1 3 Charopidae and a difference in areas of species abun- dance. In part, the apparent greater amount of anatomi- cal data for the Charopidae is illusory because many taxa were represented only by pulled fragments. This reflects the method of processing and specimen storage used at the BPBM (see Solem, 1976b, p. 19). Most fre- quently, a break occurs so that the digestive gland and ovotestis do not pull out of the shell, but sometimes only the head and terminal genitalia will be extracted. Thus, in only 21 of the 43 taxa do I have data on the ovotestis structure. For 14 additional taxa I could re- cord data on the postapical genitalia and the entire pallial region. For four taxa most of the terminal genitalia was extracted, but only part or none of the pallial region, and for four taxa it was possible to re- cord data on penial structure only. Only partial review of the anatomical patterns is possible. The sequence of discussion follows that used for the Endodontidae (Solem, 1976b, pp. 72-99). Expla- nations and data presented there are not repeated, al- though fully cross-referenced. In certain cases, refer- ence is made to extralimital taxa that have bearing on subfamily assignments. Where appropriate, their anatomy is illustrated here to provide standards against which to measure the variations found in the Pacific Island taxa. Illustration of the anatomy for new Australian and Lord Howe Island taxa will be pre- sented elsewhere, although data from them are used to establish subfamily limits. GENITAL SYSTEM The organs are discussed in approximate order from apex of the coiled visceral hump to the atrium. The terminology is modified from that originated by H. B. Baker (1938b, pp. 6-10, 92) and follows exactly the usage in Solem (1976b) with a few additions for new structures. OVOTESTIS (G) The hermaphroditic gland or ovotestis in the Charopidae normally consists of one or two clumps of palmately clavate alveoli that usually lie buried in the digestve gland above the stomach- intestine junction. The only Pacific Island species that departs from this pattern is Discocharopa aperta (M611- endorff, 1888) (fig. 34a), where the bilobed ovotestis (G) lies alongside the stomach rather than apically in the digestive gland. In Sinployea (figs. 39b, g, 43c, 53g, 57b, 60b) there may be one or two lobes, basically de- pending on exactly where the last branching of the hermaphroditic duct occurs. Generally the branching of the ducts is simple, but in Graeffedon graeffei (Mous- son, 1869) (fig. 87d) there are lateral short branches off the individual alveoli after the initial bifurcation. In Ba humbugi (fig. 75c) the ovotestis is shortened consid- erably. Jokajdon, Kubaryiellus, andLadronellum have one clump, Palline and Trukcharopa, two clumps typical differences are seen in Figure 90f, i. Species of Semperdon (fig. lOOc, g) have one or two clumps. In every case these clumps lie essentially parallel to the plane of coiling and extend apically with only an initial angling outward from the parietal wall where the start of the hermaphroditic duct lies. This is a very different pattern from the perpen- dicular to slanted orientation and division into a larger number of lobes seen in the Endodontidae (Solem, 1976b, p. 74, fig. 44). Similar orientation and multipli- cation of ovotestis lobes is found in some Australian and New Zealand Charopidae, but among the Pacific Island Charopidae and Endodontidae genera the ovo- testis differences are diagnostic. HERMAPHRODITIC DUCT (GD) As in nearly all pulmonates, this duct serves to transport sex prod- ucts from the mass of the ovotestis to the fertilization area below the expanded stomach. Its length is directly correlated with the length of the stomach, which, in turn, is dependent upon the whorl count and cross- sectional whorl area in this part of the visceral hump. In Ba humbugi, which averages only 3% whorls and has a comparatively wide aperture (fig. 74b), the her- maphroditic duct (fig. 75c) is very short compared with most other taxa. Whether the ovotestis starts after union of the ovotestis collecting tubules with a sudden or gradual expansion appears to be possibly seasonal. In no species did I have samples from the same popula- tion taken in different months of the year available in order to check whether swelling is a sign of reproduc- tive activity. None of the Charopidae show the "kink- ing" of the hermaphroditic duct seen in the Endodont- PATTERNS OF MORPHOLOGICAL VARIATION 25 idae (Solem, 1976b, p. 75, fig. 45). The charopids do show a characteristic purple-red, iridescent sheen to the hermaphroditic duct that is not seen in the en- dodontids. TALON (GT) and CARREFOUR (X) The talon (GT) varies from a simple swelling just above the en- trance of the hermaphroditic duct (GD) (fig. 39e) to a circular head on a stalk (fig. 103e). The carrefour (X) is a slight swelling in the tract after reflexion of the duct. The same pattern holds in all the species examined from the Pacific Islands. As usual in both the Charopidae and Endodontidae, there is a partial reflex- ion of the hermaphroditic duct before it enters the car- refour. This permits the compaction of apical organs when the animal has retracted into the shell. The gen- eral pattern of a circular head on a short stalk in the Pacific Island charopids contrasts with the more slen- der head pattern in the endodontids (Solem, 1976b, p. 85, fig. 49a; p. 373, fig. 164b; p. 471, fig. 199c), but the differences are small in many cases, and the variation in extralimital charopids is extensive. In both families the collecting ducts of the albumen gland empty into the carrefour below the level of the hermaphroditic duct entrance (fig. 39e; Solem, 1976b, p. 471, fig. 199c). No histological studies have been made of structures in this area. ALBUMEN GLAND (GG) The distance between the apex of the pallial cavity and the base of the stomach is less in the Charopidae than in the Endodont- idae, since the median whorl counts are, respectively, 4Va and 5V2 + . In the Endodontidae, there is more of a tendency for the albumen gland to be elongated and to occupy only part of this area. In the Charopidae, the albumen gland is almost shapeless, roughly globular, with the surface deeply indented by intestinal loops, head of the spermatheca (S), esophagus, arteries, and stomach base. Rarely does the albumen gland dissect out intact, and in order to check the shape of the talon and carrefour plus the point of entrance for the her- maphroditic duct, some of the albumen gland tissue has to be teased away before illustration. Thus, only rarely can the actual shape be shown. The size of the alveoli appears larger in the Charopidae than in the Endodontidae, but no quantification of this was at- tempted. PROSTATE (DG) and UTERUS (UT) In the En- dodontidae (Solem, 1976b, pp. 77-78; p. 373, fig. 164d; p. 471, fig. 199c) the prostate and uterus are completely separate tubes that are only lightly bound together by connective tissue. In the Charopidae they are partly (some Rotadiscinae) to completely (most Charopinae) fused, sharing a common lumen, with the prostate channel a lateral outpocket (fig. 102e) into which enter the prostatic alveoli. These are much fewer in number than in the Endodontidae, longer, and not arranged in distinct rows. Generally the prostate is as long as the expanded uterine upper section and extends slightly down along the even more expanded uterine lower chamber. The latter tapers into the free oviduct (UV) without clear demarcation, so the exact relationship between prostate and uterine length remains to be de- termined by histological studies. The uterus has a clear separation into an upper (UT 1 ) thin-walled narrower chamber and a lower (UT 2 ) thick-walled, more expanded chamber (fig. 43c-d). Presumably, actual egg encapsulation takes place in the lower chamber. This is simpler than the situation in the Endodontidae (Solem, 1976b, p. 449, fig. 191b) where four sections could be distinguished in the uterus. Comparative studies on the function and his- tology of this area in both families would be quite worthwhile. Solem (1972b, pp. 108-112) reviewed the pattern of fused versus separated pallial gonoducts in the Pulmonata and concluded that the separated condi- tion found in the Endodontidae was more primitive than the fused condition found in the Charopidae and many European-North American families. TERMINAL MALE GENITALIA The organs that may be present in this complex are the vas defer- ens (VD), epiphallus (E), penial retractor muscle (PR), penis (P), penis sheath (PS), atrium (Y), and internal structures of the epiphallus and penis proper. The lat- ter regions are used in species recognition by the snails, so that highly significant variations can be found within genera, particularly under conditions of sympatry. The discussion of variations in the Endodont- idae (Solem, 1976b, pp. 78-83) should be read as background information on the general patterns before attempting to deal with the complexities in the Charopidae, where few of these organs are constant in shape. The vas deferens (VD) starts from the end of the prostate alveoli attachment and tapers for a variable distance before becoming a narrow tube that reflexes at the penioviducal angle to ascend the penis. The area of tapering can be very short as in Russatus (fig. 90b), long in Ladronellum (fig. 102a), extend nearly to the penioviducal angle as in Tuimalila (fig. 79b), or highly variable as in Sinployea (fig. 67a, e-f, h). The ascend- ing arm of the vas deferens is without unusual features until it reaches the next structure, which differs from subfamily to subfamily of the Charopidae. In the Charopinae the vas deferens (VD) passes into an epiphallic enlargement that may be apicad or anterior of the penial retractor muscle insertion. In Charopa coma (Gray, 1843) (figs. 9b, d, 10) there is an enlargement of the vas deferens about one-third of the way from the penial retractor muscle insertion. Inter- nally (fig. 10) it can be seen that the actual differentia- tion occurs some distance after the swelling has begun. Internal pilasters of the vas deferens (VD) enlarge, terminating in finger-like projections that surround a central cavity. Below this point, which is the start of the epiphallus, two high pilasters and a lower ridge line the inside of the epiphallic chamber down to the point of entry into the penis chamber (fig. 9d). The IE PC FIG. 9. Anatomy of Charopa coma (Gray). Waiwera-Puhoi Road, north of Auckland, North Island, New Zealand. A. Solem! 11-10-1962. FMNH 135420: a, left side of foot showing size relation to shell; b, genitalia; c, details of pallial genitalia; d, internal structures of penis and lower female tract; e, pallial region; f, details of pneumostomal area; g, bottom and top views of anterior digestive system. Scale lines as marked. (PS). 26 PATTERNS OF MORPHOLOGICAL VARIATION 27 FIG. 10. Interior of lower vas deferens and epiphallus in Charopa coma (Gray). Waiwera-Puhoi Road, north of Auckland, North Island, New Zealand. A. Solem! 11-10-1962. FMNH 135420. Greatly en- larged. (NB). penis retractor muscle (PR) inserts as a U-shaped fan around the lower part of the epiphallus (E). Various modifications occur extralimitally in such taxa as Phenacohelix pilula (Reeve, 1852) (fig. 11). Here the penial retractor muscle inserts on the head of the epiphallus, and the vas deferens enters almost later- ally. The shift in muscle attachment is relatively insignificant, but does make external recognition of the epiphallus presence difficult if not actually impos- sible. The Pacific Island Charopinae mostly show, in- sofar as it could be checked, a specialized vas defer- ens-epiphallus junction (figs, 57e, 79c). Exceptions are found in Graeffedon and the provisionally assigned Discocharopa (see below). A circular sphincter sur- rounds the epiphallic pore (DE), and a Y-shaped plug extends through the pore to attach onto the wall of the epiphallus as a circular ridge. In general, the walls of the epiphallus are without sculpture or have the typi- cal ridges (fig. 87c) seen in Charopa coma (Gray, 1843) (fig. 10). Because of size and preservation problems, only a few of the species could be checked for this structure, and illustrations are given only for Sinployea complementaria (Mousson, 1865) (fig. 57e) and Tuimalila pilsbryi (fig. 79c). It was observed in several other Sinployea. Normally the head of the epiphallus expands abruptly, indeed it can be bulbous (fig. 67a), so that external recognition is easy. In the Semperdoninae (figs. lOOb, e-g, 102a, 103a, c-d), the vas deferens remains a thin tube during its ascent and enters laterally into the epiphallus, with the penial retractor muscle inserting directly onto the head of the epiphallus. The exact point of entry of the vas deferens into the epiphallus differs. In Himero- concha and Ladronellum the entry occurs slightly below the rounded head (fig. 102d), whereas in Sem- perdon the entry is right next to one edge of the penial retractor muscle insertion (fig. lOOi). The epiphallus is a double-walled tube, sometimes coiled within the muscle sheath. The outer wall is a thin muscle sheath, and the inner a thick glandular tube usually rolled inward on one side. In the Trukcharopinae (figs. 90b-d, f-g, i-j, 93b-c, f, h) the vas deferens enters directly into the penis through a simple pore. In all examined genera except Jokajdon (fig. 93b-c) the vas deferens passes through the muscle fan before entering the penis. In Jokajdon the muscle insertion is directly on the head of the penis, and the vas deferens enters lateral to the inser- tion. This is interpreted as a secondary modification probably resulting from reduction in size. The Truk- charopinae thus have no trace of an epiphallus. In the Rotadiscinae (H. B. Baker, 1927, pi. 16, fig. 14; pi. 17, figs. 22, 27) the epiphallus is less clearly differentiated externally, and it is either before (Radioconus) or after (Rotadiscus, Radiodiscus) inser- tion of the penial retractor muscle. The enlarged pilas- ters of the epiphallus are the main differentiating fea- tures. In all of the Pacific Island Charopidae examined to date the penial retractor muscle originates from the diaphragm near the apex of the pallial cavity and in- serts onto the penis or epiphallus. There is no shift to the columellar muscle equivalent to that seen in the Endodontidae (Solem, 1976b, pp. 81-83), which per- mitted elongation of the penis in such taxa as En- dodonta and Australdonta. In addition, generally the penial retractor muscle in the Charopidae is very short, at times scarcely a tuft connecting the dia- phragm and penis or epiphallus head. This is another correlative of the reduced whorl count in the Charopidae as compared with the Endodontidae. With a few exceptions, the external appearance of the penis in the Pacific Island Charopidae is for a bul- bous apical section that then tapers either abruptly or quickly to a slender stalk of variable length that joins the vagina to form the short atrium. The degree of the apical bulge depends directly upon the size and com- plexity of the internal pilasters and vergic structures. Figure 67 gives a fair sample of shape variations, and KD GD FIG. 11. Anatomy of Phenacohelix pilula (Reeve). Church Road, Kaitaia, Northland, North Island, New Zealand. L. Price! X-1962. FMNH 135421: a, pallial region; b, extended foot showing caudal horn (CH) and pedal grooves (FS); c, genitalia; d, junction of hermaphroditic duct (GD) and talon (GT); e, internal structure of penis. Scale lines as marked. (PS). 28 PATTERNS OF MORPHOLOGICAL VARIATION 29 this is within a single genus. The major exception is the Semperdoninae (figs. 100, 102, 103) where a penis sheath, narrowed collar between epiphallus and penis, plus a quite different pattern of internal structure have combined to produce a more tubular appearance. It is not possible to assign a species to a genus on the basis of the external penis appearance. The internal struc- tures must be examined. They differ radically from subfamily to subfamily, and within a genus are subject to extensive modifications for species recognition pur- poses. A review of structure on a subfamily by subfam- ily basis is required. In the Charopinae from Australia, New Zealand, New Caledonia, Lord Howe Island, South Africa, St. Helena, and South America there is an amazing amount of variation that will have to be dealt with elsewhere. Comments here will be restricted to the type genus, Charopa Albers, 1860, Phenacohelix Suter, 1892, and the Pacific Island taxa. Charopa coma (Gray, 1843) has the penis with only a modestly swollen ex- terior (fig. 9b) and a very short basal stalk portion. Internally (fig. 9d) it shows an apical verge (PV) with near terminal slitlike pore, below this is a pocket stimulator (PC), followed by a series of variable-sized circular pilasters that occupy a quite low position. In addition, the upper and middle walls of the penis chamber have raised, soft pustules. Phenacohelix pilula (Reeve, 1852) has a much elongated and more slender penis (fig. lie), with shifted position of the epiphallus (E) to below the insertion of penial retractor muscle (PR). Internally (fig. lie) a small verge (PV) with terminal pore lies above a series of corrugated longitudinal pilasters and a vague swelling that I interpret as a remnant of the pocket stimulator. The verge and ridge correspond to the external bulge on the penis section. The sculpture of the wall, position and number of circular ridges, size of pocket stimulator, and verge size differ between Phenacohelix and Charopa, but the fact of a basic set of structures is the important data. In most of the Pacific Island Charopinae there is simple variation in the relative sizes of the verge, pocket or modified pocket stimulator, and the number and prominence of the circular bands. The wall sculpture found in the New Zealand taxa seems to be absent in the Pacific Island taxa. The combination of very small size (see table XV), limited material, and preparation of illustrations at an early stage in the study means that treatment of the variations in this portion of the study is inadequate. Specific comments on differences between geographically close species are given in appropriate places in the systematic review. Here I will mention only a few more obvious changes. The verge varies from a small spade-shaped papilla with open side (fig. 42e, S. peasei), to a conical projec- tion with terminal pore (fig. 43e, S. auanaensis), or it is reduced to a swelling attached to one wall (fig. 55e, S. allecta), or a globular swelling (fig. 57c, S. com- plementaria, fig. 75h, Ba humbugi), a protruding lobe with laterally apical pore (fig. 79c, Tuimalila pilsbryi), or the bulbous structure seen in Vatusila tongensis (fig. 84b). The circular ridge can be split into several and low (fig. 42d-e, Sinployea peasei), doubled and thick (fig. 43e, S. auanaensis), single and thin (figs. 51h, S. inter- media; 57c, S. complementaria; 67j, S. euryomphala), or very thick and partly altered into pads (figs. 55e, S. allecta, 79c, Tuimalila pilsbryi). The pocket stimulator is particularly variable and with few exceptions has been inadequately illustrated and interpreted. Comparison of the doughnut shape in Figure 39h, split "U" in Figure 42d, raised globular mass in Figure 53c, low "U" in Figure 53h, lobular structure in Figure 55e, "U" ridge in Figure 67d, widely open pocket in Figure 67j, greatly reduced and altered structure in Figure 75h, and nearly closed pocket in Figure 79d gives an idea of the great extent of variation in this structure, but this is only an introduc- tion to its complexity. This study demonstrates that extensive variation occurs in verge, circular ridges, and pocket stimulator in the Pacific Island Charopinae, but analysis of the patterns of variation is beyond the scope of this report. Because Sinployea is evidently in a highly active stage of speciation, study of this area might yield important data, but is left for others. A major modification of this pattern occurs in Graeffedon (fig. 87c). The epiphallus (E) opens through a simple pore that is surrounded by a huge circular pilaster (PP). Immediately below this are three large pilasters: an upper, free-tipped vergic stimulator, a long and low semicircular ridge oriented longitudi- nally in the penis chamber, and a broad, low ridge with slightly free upper tip that occupies the lower half of the penis. Coupled with the absence of the valve at the vas deferens-epiphallus junction, these structures mark Graeffedon as remote from the other Charopinae, but it is better accommodated in this subfamily than assigned to one of its own. The Trukcharopinae (fig. 90c-d, g, i) have a simple pore opening from the epiphallus (Russatus), a low vergic papilla (Kubaryiellus), or even a tubular verge (Palline). Various pilasters and a possibly highly mod- ified pocket stimulator are found in some of the taxa. These variations are discussed more fully on pp. 205- 207. Many of these represent additive structures com- pared with the Charopinae, but basically they could be derived from some of the Charopinae conditions in ex- tralimital taxa (Solem, unpublished data). The Semperdoninae (figs. lOOd, f, j, 102d, 103b, f) have the epiphallus entering the penis through a nar- row muscular collar, with high lamellar pilasters ex- tending down and then usually coalescing into three, high glandular pilasters. The whole penis is sur- rounded by a muscular sheath. Exceptions to the basic pilaster pattern are Himeroconcha rotula (Quadras & Mollendorff, 1894), which lacks the initial radiating pilasters although retaining the three basal ones, and Ladronellum mariannarum (Quadras & Mollendorff, 30 SOLEM: ENDODONTOID LAND SNAILS 1894) in which the initial radiating pilasters coalesce into a huge horseshoe-shaped pilaster (fig. 102d) that is inflatable with fluids. The enclosure of both epiphallus and penis by a muscle sheath is as great a change as is the pilaster pattern. These make the Semperdoninae the most isolated of the charopid subfamilies in terms of penis structure. In the few dissected Rotadiscinae (H. B. Baker, 1927; Solem, unpublished data) the epiphallus enters through a pore, papilla, or short verge into a very thick-walled chamber with longitudinal pilasters. There may or may not be accessory organs associated with the penis and/or atrium. This represents yet another series of experiments in the Charopidae. Finally, the situation in Discocharopa (fig. 34c) re- quires comment. A short, tubular penis with lateral entrance of the vas deferens has the penial retractor muscle inserting apically. Apparently there are at least two longitudinal pilasters inside the penis, but the available material was not adequate to work out the details (see p. 75). Although this departs from the reported pattern for the Charopinae (see above), the genus is temporarily placed in the Charopinae pending revision of the extralimital groups. It probably will be split into another subfamily unit, but available data are inadequate to propose a more appropriate classifi- cation. The atrium, in all dissected material, is a simple, short tube opening externally behind and above the right rhinophore. It varies somewhat in length but shows no really significant changes. TERMINAL FEMALE GENITALIA The post- uterine or free oviduct (UV), spermatheca (S), and va- gina (V) in the Endodontidae are very slender tubes without unusual structures. They show variation pri- marily where the spermathecal shaft inserts penis, atrium, or free oviduct (Solem, 1976b, pp. 83-84). In the Charopidae the situation is quite different. Nor- mally, the spermathecal base and at least parts of the vagina are greatly enlarged and with complex internal pilasters. The free oviduct, most of the vagina, and the spermathecal shaft above the swelling will be quite slender, with an ovate, expanded head of the sperm- atheca buried in the albumen gland-prostate margin. The shape and proportions shown by Sinployea avan- aensis (fig. 43c) are typical. The shifted expanded area in S. aunuuana (fig. 53b) or expanded free oviduct (UV) in S. allecta (Cox, 1870) (fig. 55c), shortened va- gina in S. inermis lakembana (fig. 67f), and generally greater elongation in the Semperdoninae (figs. lOOb, e, g, 102a, 103a, c-d) are the typical minor variations. The nearest thing to a major variation is found in Jokajdon (fig. 93bc) where the expanded area is re- stricted to the base of the vagina and atrium, and the normally expanded areas are slender in comparison. In addition there is an accessory muscle attaching to the penioviducal angle. These changes may correlate with the very restricted shell aperture and large barriers of Jokajdon. Internally there appear to be two basic patterns of sculpture of the upper vagina and lower spermathrca. In the typical Charopinae, such as S. complementaria (Mousson, 1865) (fig. 57f), S. allecta (Cox, 1870), and Ba humbugi, there are weak longitudinal pilasters in the vagina, the free oviduct (UV) has a constricting pilaster with central pore (UVO), and the opening of the spermatheca (S) has a central pore (SO) through a circular pilaster with one edge a free flap. Because of size problems and the early stage in the study at which this area was examined, only the one has been illus- trated. The second pattern is seen quite clearly in Graeffedon (fig. 87c) where the opening to the free oviduct is a simple pore and there are huge longitudi- nal pilasters lining the walls of the spermatheca and vagina. Essentially the same pattern is seen in Palikirus, Palline, Semperdon uncatus, S. hepta- ptychius (Quadras & Mollendorff, 1894), Ladronellum mariannarum (Quadras & Mollendorff, 1894), Himeroconcha lamlanensis, and H. fusca (Quadras & Mollendorff, 1894), thus indicating that this general type occurs not only in some of the Charopinae, but also in the Trukcharopinae and Semperdoninae. A de- tailed comparative study of this region would be --veil worthwhile but is beyond the scope of this monograph. PALLIAL COMPLEX A discussion of the functional significance that the closed and complete ureter found in most Charopidae has in relation to water conservation as contrasted to the incomplete posteriorly opening ureter of the En- dodontidae was given previously (Solem, 1976b, pp. 84-87). Some of the New World Rotadiscinae show only a partial closed ureter (H. B. Baker, 1927, pi. 16, fig. 17, pi. 17, figs. 24, 30), and there are Australian equivalents (Solem, unpublished data). All of the Pacific Island Charopidae that have been dissected show a complete ureter with the ureteric pore opening next to the anus just inside the pneumostome. This contrasts immediately with the Endodontidae where the ureteric pore opens at the posterior part of the pal- lial cavity near where the anterior margin of the rectal kidney lobe touches the hindgut (Solem, 1976b, p. 85, fig. 49c). Thus, even a glance at the anterior part of the pallial cavity is sufficient to tell whether a Pacific Is- land species belongs to the Endodontidae or Charopidae. A typical charopid configuration is seen in Sinployea allecta allecta (Cox, 1870). Viewed from the outside (fig. 55a), the bilobed nature of the kidney (K) is obvious, with the longer, cigar-shaped rectal lobe (lower in figure) definitely overlapping the hindgut (HG). The shorter, irregularly shaped pericardial lobe (upper in figure) partly overlies and is cupped partly around the heart (H). The kidney base (left in figure) stops short of the downward twist of the hindgut as intestine and is abutted by the lobules of the digestive gland. The primary ureter (KD) originates from near the anterior margin of the pericardial lobe of the kid- PATTERNS OF MORPHOLOGICAL VARIATION 31 ney, follows the upper margin of this lobe posteriorly, lying partly on the pallial cavity roof and partly on the kidney itself, reflexes abruptly and as the secondary ureter follows the lower margin of the rectal kidney lobe anteriorly to its termination, then lies next to the hindgut until both disappear under the mantle collar (MC). The principal pulmonary vein (HV) extends an- teriorly from the heart along the pallial roof, but in the smaller species shows no sign of branching, fading out from visual observation well short of the mantle collar. The bilobed kidney, or in a torn and extracted specimen, the tip of the rectal lobe of the kidney, plus the presence of the secondary ureter as a distinct tube next to the hindgut, are sufficient to immediately iden- tify a species as a member of the Charopidae or Punctidae rather than the Endodontidae. In the latter family (Solem, 1976b, p. 459, fig. 195a) there is at most a short rectal lobe with the ureter ending in a ureteric pore (KX) as it reaches the hindgut. In a few Charopidae, such as Phenacohelix pilula (Reeve, 1852) (fig. lla), the rectal arm is as short as in the Endodon- tidae, but the clear presence of the secondary ureter along the hindgut is sufficient for family separation. Exceptions to this in some Australian taxa will be con- sidered elsewhere. When dissected out (fig. 55b) and viewed from an inside view of the pallial cavity, the typical charopid pallial region shows only minor additional features. Remnants of a retractor muscle to a rather strongly developed muscle come off the parietal-palatal margin near the kidney base. A weaker version of the same muscle can be seen in many species of Endodontidae (Solem, 1976b, p. 459, fig. 195a). The heart in this view clearly lies on top of the kidney, the rectal kidney lobe extends partly under the hindgut, thus overlapping on both sides of this tube, and both the hindgut and ureter terminate just inside the mantle collar. Quite possibly this area is involved in water resorption and is a pre- cursor of the more elaborate structures seen in higher aulacopods, such as an apparent bladder in Deroceras reticulatum (Miiller, 1774) (Runham & Hunter, 1970, pp. 77-79, figs. 32-33). Modifications of this general pattern involve addi- tion of new structures, changes in the relative size of the kidney lobes, differences in the amount of space between the two lobes, and compactional alterations correlated with reduction in whorl counts. Data on at least part of the pallial complex was recorded for 36 of the 43 taxa for which at least some anatomical mate- rial was available. Illustrations are presented for 20 of these. The only additive structure seen in a Pacific Is- land taxon is the extensive intrusion of mantle gland tissue onto the pallial roof in Graeffedon graeffei (Mousson, 1869) (fig. 87a, MG) and a slight extension in Semperdon xyleborus. This phenomenon is far more common in extralimital taxa, with the intrusion rang- ing from a massive and sharply defined area in Champa coma (Gray, 1843) (fig. 9a) to the short area in Phenacohelix pilula (Reeve, 1852) (fig. lla) and the elongated finger in Laoma leimonias (Gray, 1850) (fig. 25a). The function of this extension is unknown. Although there is a modest amount of intraspecific variation in the relative length of the two kidney lobes, much of this seems to be caused by differential contrac- tion and compaction when the animal withdraws into the shell. Dissections made from deeply retracted specimens tended to have the rectal kidney lobe un- changed in shape and position, but pulled back further relative to the heart and pericardial kidney lobe. The latter would tend to be somewhat twisted and distorted (for example see fig. 34d). I made no exact measure- ments as to the relative lengths because the results would not be strictly comparable. Table IX summarizes the relative lengths in the Pacific Island species. Only in Discocharopa aperta (Mollendorff, 1888) (fig. 34d) is the rectal lobe somewhat reduced to as great an extent as in the New Zealand Phenacohelix pilula (Reeve, 1852) (fig. lla). The general pattern is for the rectal lobe of the kidney to be much longer than the pericar- dial. This reaches its greatest extent in such taxa as Graeffedon graeffei (Mousson, 1869) (fig. 87a) and Semperdon heptaptychius (Quadras & Mollendorff, 1894) (fig. lOOa). In both of these taxa the pericardial lobe is reduced to a small fraction of the length and volume of the rectal. The length relationship varies within a genus, since in Sinployea there are nine species in which the TABLE IX. - RELATIVE LENGTH OF KIDNEY LOBES IN THE PACIFIC ISLAND CHAROPIDAE. Subequal or Equal Sinployea tahitiensis ^. lamellicosta j>. avanaensis ^. intermedia S^. allecta allecta ji. vicaria vicaria ^. kusaieana j>. inermis inermis J[. i_. lakemba ^. adposita Ba humbugi Tulmalila pilsbryi Russatus nigrescens Himeroconcha fusca H. lamlanensis Rectal Much Longer Sinployea modicella ^>. montana S, neglecta S. peasei S^. aunuuana ^. clista S_. complement aria S^. ir_regulaj^is_ S^. eu r y ompha 1 a. Graeffedon graeffei Kubaryiellus kubaryi Trukcharopa trukana Palikirus cosmetus Jokajdon calllzonus Palline np_t era not era P_. micramyla Semperdon uncatus j^. xyleborus ^. heptaptychius S_. rotanus Ladronellum raariannarum Pericardial Much Longer Discocharopa aperta 32 SOLEM: ENDODONTOID LAND SNAILS kidney lobes are equal or nearly equal in length and nine in which the rectal is significantly longer. I could detect no conchological correlations in size, shape, or whorl count with this variation in kidney lobe length concerning Sinployea. The other Charopinae have a secondarily shortened and very fat kidney correlated with whorl count reduction (Ba humbugi, fig. 75a), subequal lobes in the gigantic Tuimalila pilsbryi (fig. 79a), or the very elongated rectal lobe in Graeffedon graeffei (Mousson, 1869) (fig. 87a). In the Truk- charopinae only Russatus nigrescens (Mollendorff, 1900) (fig. 90a), whose whorl count is reduced to a mean of 3Vs + , has a shortened kidney, whereas in the Semperdoninae only the large-sized Himeroconcha have the kidney lobes equal or subequal in length. In both the Trukcharopinae and the Semper- doninae the arms of the ureter are tightly pressed to- gether without any pallial roof tissue visible between them. The pattern is somewhat unusual in Russatus nigrescens (Mollendorff, 1900) (fig. 90a), with the abrupt angling of both ureter arms at the anterior margins of the kidney lobes. I interpret this as the result of shortening and secondary thickening of the kidney in relation to whorl and pallial cavity reduc- tion. In the Charopinae, the same compaction is seen in Ba humbugi (fig. 75a), with only slightly less abrupt and extensive angling of the ureter. Of the examined Sinployea, all taxa except S. allecta allecta (Cox, 1870) (fig. 55a-b) and S. irregularis (Garrett, 1887) have at least a narrow strip of pallial roof tissue visible be- tween the ureter arms, as does Tuimalila pilsbryi (fig. 79a). Extralimital Charopinae are variable in this fea- ture (see figs. 9e, lla). Shortening of the pallial cavity is more common than elongation. The latter seems to have occurred only in Jokajdon callizonus (Mollendorff, 1900) (fig. 93a) where it extends % of a whorl apically. This prob- ably correlates with the narrowed cross-section of the body whorl in this species (fig. 92b). Major shortening has occurred in Ba humbugi (fig. 75a) and Russatus nigrescens (Mollendorff, 1900) (fig. 90a). Their respec- tive mean whorl counts of 3% and 3Vs+ are the lowest in the Pacific Island Charopidae, except for the rotadis- cine Microcharopa mimula (mean whorl count 3V4-). In both Ba and Russatus the width of the kidney is almost equal to its length, and the arms of the ureter are tightly compacted and overlap each other between the kidney lobes. The process has been carried further in Russatus, with V4 whorl length to the pallial cavity, than inBa, where the length is still Vz whorl. The same type of compaction has occurred in both genera, and they form a contrast to the situation in the New Guinea charopids Pilsbrycharopa and Paryphantopsis (Solem, 1970a, p. 250, fig. 2a, f) in which kidney com- paction resulted in progressive increase of the angle between the arms of the ureter, and the kidney is ro- tated away from the hindgut. Yet another pattern is seen in the New Zealand flammulinid and Maori- concha groups. The pallial cavities in the Pacific Island Endodon- tidae and Charopidae have radically different ureter and kidney structures. Within the Pacific Island Charopidae there is a relatively simple pattern of vari- ation in relative lengths of the kidney lobes and the extent to which lung roof tissue is visible between the arms of the ureter. DIGESTIVE SYSTEM Allowing for the generally reduced whorl counts (table X) in the Charopidae, the gross features of the endodontid (Solem 1976b, pp. 372-373, figs. 163-164) and charopid (figs. 9e, g, 75a) digestive tracts are the same. The digestive glands (OG) are in contact above the esophagus (BE), which continues past the pallial cavity as a slender tube. In some taxa, such as the larger Sinployea, Ladronellum mariannarum (Mollen- dorff, 1900), and Kubaryiellus kubaryi (Mollendorff, 1900), the salivary glands are fused above the esophagus, whereas in Ba humbugi, they do not even touch posteriorly. I am not certain if there is any sys- tematic significance to these changes. Intestinal loops above the pallial cavity apex occupy about Vie of a whorl, instead of the Vs whorl in the Endodontidae. The charopid stomach expansion averages about % of a whorl, with the digestive gland shorter and more com- pact than in the Endodontidae. Except for these changes associated with the reduced whorl count, the gross features of the digestive tract are the same in the two families. Radular features are one of the best guides to separating the Endodontidae, Charopidae, and Punc- tidae on the Pacific Islands. The endodontid radula (Solem, 1976b, pp. 88-94, figs. 51-54) has a tricuspid central, several bicuspid laterals, and marginals that have the ectocone fragmented and the endocone in- creasing almost to the size of the mesocone. There are very noticeable differences in the angling of the indi- vidual teeth, but their small size and difficulties in mounting and preparation have prevented full study of their variation. The punctid radula (fig. 12) presents a number of unusual features. The central tooth (fig. 12a, upper) has a long slender mesocone and two very slender, much shorter ectocones. The laterals (fig. 12b-e) have two slender cusps and three much shorter accessory cusps, whereas the outermost lateromarginals (fig. 12f) are broader and with the outer large cusp reduced in size. The accessory cusps are at to below the limit of optical microscope examination, depending upon the quality of the equipment and illumination. As far as I am aware, the only optical microscopist to detect and illustrate these accessory cusps was H. B. Baker (1927, pi. 16, fig. 11), also reprinted in Pilsbry (1948, p. 642, fig. 349, d). In addition to the obvious cusp edges and numbers there are fundamental differences in the nature of the cusps and basal plates in the two families. The en- dodontid basal plate (Solem, 1976b, p. 89, fig. 52d) has a typical interlocking relationship with the next pos- FIG. 12. Radular teeth ofPunctum minutissimum (Lea). Cedar bog, Woodburn Road, 4 miles southwest of Urbana, Champaign County, Ohio. E. Keferl! X-20-1969. FMNH 151102: a, central (upper) and 1st lateral (23,000x); b, early lateral teeth (10,800x); c, single midlateral tooth (21,500x); d, late lateromarginal teeth (15,900x); e, low angle view of lateromarginal teeth (10,700x); f, outer lateromarginal tooth showing cusp reduction (28,000x). 33 34 SOLEM: ENDODONTOID LAND SNAILS terior tooth, is square to rectangular in shape (Solem, 1976b, p. 91, fig. 54b, e), and, at least for the early lateral teeth, the cusps are sharply pointed and ele- vated at a high angle (Solem, 1976b, p. 88, fig. 51a-b; p. 89, fig., 52a-b, d). In the Punctidae (fig. 12c, e) the basal plate is long and tapering, has no apparent inter- locking relationships, and the cusps are bluntly rounded and point almost directly forward. It is unfor- tunate that these differences can be seen only with the scanning electron microscope, a fact that reduces their routine utility in identification and classification. The Pacific Island Charopidae have a very stan- dard pattern of structure (figs. 13-14). As pointed out earlier (Solem, 1976b, p. 93), extralimital charopids show a variety of structure, but review of these is out- side the scope of this monograph. The tricuspid central tooth (figs. 13b, 14a) is slightly narrower and shorter than the adjacent laterals, which are tricuspid. When viewed at a low angle, it is evident that the side cusps of both central and laterals are raised above the eleva- tion plane of the mesocone (fig. 14a), whereas in the Endodontidae (Solem, 1976b, p. 88, fig. 51a) they are in the same elevation plane. The outer laterals and early marginals in the Charopidae (fig. 14b) do not have the elevated side cusps. The marginals may retain the tricuspid pattern (fig. 13c-f), or rarely the outer mar- ginals may become multicuspid. Climo (1969a, figs. 31-34; 1970, figs. 11-14) gives a number of radular transects for New Zealand charopids. Because only a few species could be examined with the SEM and their pattern of structure was quite uniform, results from optical viewing mostly have been omitted from the text in this volume. The bicuspid laterals of the Endodontidae, tricus- pid laterals of the Charopidae, and multicuspid lateromarginals of the Punctidae found on Pacific Is- lands thus present clearcut differences among the families. The Austro-Zelandic charopids, however, show a great variety of radular structures, so that the endodontid-charopid distinction does not hold for that area. Jaw structure in the Charopidae was not studied in detail because the pattern of partial plate fusion in larger taxa paralleled the situation found in the En- dodontidae (Solem, 1976b, p. 94). FREE MUSCLE SYSTEM As in the Endodontidae (Solem, 1976b, p. 94), all dissected charopids had the right ommatophoral re- tractor passing through the penioviducal angle and joining the right rhinophoral retractor that passes out- side the penioviducal angle posteriorly. Unlike the En- dodontidae, in all examined Charopidae the penial re- tractor muscle arises from the diaphragm and inserts onto the penis or epiphallus. Only a few alterations in the common pattern of unions were observed. Russatus nigrescens (Mollendorff, 1900) has the tentacular re- tractors fusing with the tail fan at the columellar muscle rather than earlier as in most taxa. InJokajdon callizonus (Mollendorff, 1900) a new muscle attaches at the penioviducal angle, joining the tail fan much later. With the elongated (% whorl) pallial cavity and nar- rowed aperture with large barriers (fig. 92b) in this species, the added muscle may play a major role in successful body retraction. In Sinployea complementaria (Mousson, 1865) there is a muscle from the right tentacular retractor that inserts on the apex of the free oviduct, whereas in Tuimalila pilsbryi a muscle runs from the columellar retractor to the apex of the free oviduct. It is possible that a weaker version of this same muscle is present in smaller species of Pacific Island Charopidae and was overlooked in earlier phases of this study. It was not possible to recheck all taxa for this feature. Extralimi- tal taxa, such asStephanoda binneyana (Pfeiffer, 1847) (fig. 31c), show a vaginal retractor muscle (VRM). It is probable that a number of such experiments in added muscles exist, but have not been observed. NERVOUS SYSTEM Wherever possible, the enervation of the penis from the right cerebral ganglion was confirmed, but because of difficulties in handling the small-sized material, preservation in alcohol, and heavy covering of connective tissue over the ganglia, no attempt at working out the details of the nervous system was made. Climo (1970, fig. 21B) illustrated the central nervous system ofPhenacharopa novoseelandica (Pfeif- fer, 1853). EXTERNAL BODY FEATURES In the Trukcharopinae and Semperdoninae, except for Semperdon heptaptychius (Quadras & Mollendorff, 1894), Ladronellum mariannarum (Quadras & Mol- lendorff, 1894), and Himeroconcha fusca (Quadras & Mollendorff, 1894), the body color is yellow-white, without darker markings. In the latter three species there are gray to reddish gray markings on the neck, ommatophores, and mantle collar. In the Charopinae, all live-collected Cook, Society, and Samoan Island species of Sinployea, except the Swains Island S. in- termedia, have light to dark gray markings. All of the Sinployea with gray markings have been taken in semiarboreal situations. The purely terrestrial S. kusaieana, S. euryomphala, S. inermis, S. adposita, and S. irregularis (Garrett, 1887) have yellow-white body color. Tuimalila pilsbryi also shows the darker mantle and neck coloration. The dark body color ap- pears to be a correlative of semiarboreal habitat. Many of the arboreal New Zealand and Australian charopid taxa have a strongly developed mucus appa- ratus at the hind end of the foot. A typical "mucus pore" or "caudal foss" (CF) is present in Phenacohelix pilula (Reeve, 1852) (fig. lib), with the pore overhung by a caudal horn (CH). This is effectively an intensifi- cation of the point where the foot grooves unite above the tail plus a mucus-secreting gland concentrating at FIG. 13. Radular teeth of Tuimalila pilsbryi. Station T-22, 1,000ft. elevation, Eua, Tonga. FMNH 152378: a, near middle of radula at 280 x; b, central tooth and 1st lateral on right side of radula at 2,820 x ; c, early laterals on right side of radula at 2,900 x ; d, transition from laterals to marginals on left side of radula at 2,950x; e, middle marginals from left side at 2,875x; f, outermost marginals at 2,875x. All views nearly vertical. 35 36 SOLEM: ENDODONTOID LAND SNAILS FIG. 14. Radular teeth ofTuimalilapilsbryi. Station T-22, 1,000 ft. elevation, Eua, Tonga. FMNH 152378: a, 45 angle view of central (c) and early laterals at 3,200 x showing pattern of cusp elevation; b, same view of transitional zone between laterals and marginals at 3,125x. this point. This apparatus is absent in Charopa coma (Gray, 1843) (fig. 9a) and all of the dissected Pacific Island Charopidae (fig. 43a). The presence of this pore is associated with arboreal snails (Solem, 1976b, pp. 105-106) and is not of major phyletic importance, as was also confirmed by Pilsbry (1892a-b, 1893a-b) and Climo (1969a, pp. 148-150). As in the Endodontidae (Solem, 1976b, p. 94), the mantle collar normally is without protrusions, the gonopore is located below and slightly behind the right ommatophore, and the slime network is rather weakly defined. Neither family shows substantial external modifications in the Pacific Island taxa. PATTERNS OF VISCERAL HUMP REDUCTION In the Endodontidae (Solem, 1976b, pp. 94-98) one of the major repetitive changes is an increase in the number of whorls and thus elongation of the visceral hump. Not only do the Pacific Island Charopidae have a much lower average whorl count than the Endodont- idae (median means 4Vfe and 5V2 + , respectively), but of the Charopidae only Semperdon kororensis (Bed- dome, 1889) with 5Vs whorls and the Rarotongan Sinployea planospira (Garrett, 1881) with 6% whorls exceed a mean count of five whorls. The latter has not been taken in this century, and of the former only fragmentary extracted pallial collars and terminal genitalia were available. Thus I can offer no data of any changes associated with elongation of the visceral hump. Reduction in whorl count and thus in the total length of the visceral hump is a common pattern in charopids from many areas of the world. Such taxa as the Juan Fernandez A mphidoxa (fig. 30a, c), New Zea- land Flammulina, Maoriconcha, and Otoconcha (Climo, 1971a), a few of the South African Trachycystis (Connolly, 1939), and some New Caledonian taxa (Solem, 1961) show varying degrees of visceral hump reduction. This is carried furthest in the New Zealand Otoconchinae and the semislug Ranfurlya. Discussion of their changes is beyond the scope of this review, but the general pattern is for zonal compaction of organs in several systems (see Solem, 1966, for a discussion of this in the Thailand Helicarionidae). In the Pacific Island Charopidae whorl reduction takes two forms simple decrease in mean whorl count without any change in whorl profile (Microcharopa mimula, Z l A-;Lagivala minusculus, 3%; Discocharopa aperta, 3%; and Palikirus ponapicus, 3%) or reduction in whorl count accompanied by a drastic increase in cross-sectional areas of the body whorl (Russatus ni- grescens, 3Vs + ',Ba humbugi, 3%). Either the former taxa have not been dissected, or, in the case of Discocharopa, there are no close relatives known with which anatomical comparisons can be made. Thus, comments here must be restricted to the latter situation, whorl count reduction combined with whorl cross-section increment. The degree of whorl changes can be judged by comparing whorl increment rates and whorl profiles in Kubaryiellus and Russatus (fig. 89a-b, d-e) and then in Sinployea irregularis (Garrett, 1887) (fig. 65d-e) and Ba humbugi (fig. 74a-b). Anatomically, the typically trukcharopinine half whorl pallial region of Kubaryiellus (fig. 90e) can easily be altered to the nearly square kidney (K) and one quarter whorl pallial cavity of Russatus (fig. 90a) by shortening and widening of the kidney, ventral flexing PATTERNS OF MORPHOLOGICAL VARIATION 37 of the intestinal loop, and a slight ventral movement of the anterial part of the pericardial kidney lobe over the heart (H) and principal pulmonary vein (HV). Compar- ing the gross genitalia (fig. 90b, f), the proportionately shortened free oviduct (UV), spermatheca (S), prostate (DG), and uterus (UT) ofRussatus are obvious. There is no detectable change in the penis (P) and vagina (V). The pallial region of Sinployea irregularis (Garrett, 1887) was not illustrated, but it has the typical generic pattern of extending % of a whorl apically, there is no lung roof space visible between the arms of the ureter, and the rectal lobe is distinctly longer than the pericardial. In Ba humbugi (fig. 75a) the anterior half of the half whorl pallial cavity has undergone shorten- ing and change, with the anterior margins of the kid- ney flared laterally, but the posterior portion of the kidney and the intestinal loops are essentially unmod- ified when compared with the degree of change seen in Russatus (fig. 90a). Contrasting the genitalia of S. ir- regularis (fig. 67a, drawn from a deeply retracted specimen) and So (fig. 75b-e, h) shows a rather drastic folding of the prostate and uterus plus altered insertion of the penial retractor muscle in the latter taxon. Ba also shows shortening of the vagina and spermatheca. Thus, changes in the visceral hump length of Pacific Island Charopidae involve selective shorten- ings in portions of the organ systems lying in the pal- lial region. Unless there is clear shortening of the neck region (area between ommatophores and pallial collar edge when animal is crawling), the prostate-uterine area and spermathecal-free oviduct sections are more apt to be involved than the penis-vagina, and the kid- ney area, than the gas exchange surfaces of the pallial roof. More extended comments on these changes are postponed pending completion of studies on Australian and New Zealand taxa. SUMMARY OF ANATOMICAL VARIATION The Pacific Island Charopidae show a number of minor variations in the terminal genitalia and kidney configurations. The former involve species recognition factors, the latter are of uncertain significance except in cases of obvious elongation or shortening of the pal- lial cavity. These terminal genitalia patterns are not continuously variable, but fall into rather discrete general patterns that indicate multiple colonizations of the island areas and are used in part to recognize sub- family units. The major and minor anatomical differences be- tween the Pacific Island Charopidae and Endodontidae were summarized by Solem (1976b, pp. 97-98, tables LVIII-LIX). Consideration of the complexities in the Australian, New Zealand, and New Caledonian taxa must be deferred. There are no clear unitary anatomi- cal trends within the Pacific Island Charopidae com- pared with those outlined for the Endodontidae (Solem, 1976b, pp. 98-99), although the addition of structures among the various charopid subfamilies is striking. A major correlative of the "looser" whorl coiling pattern (pp. 41-43) in the Charopidae compared with the Endodontidae is that the cross-sectional area of the body whorl in particular is greater in the former fam- ily. This has important implications on the size of the anterior body and its contained organs. The wider area in the Charopidae permits widening of the terminal genitalia. The much fatter penis and vagina-free oviduct-spermathecal union area in the Charopidae is possible directly because of this extra space provided by the looser coiling. Whereas in the Endodontidae these organs all are thin tubes, in most Charopidae they are thick and with complex internal structures. The only Pacific Island charopid known to have clearly narrowed structures in this area is Jokajdon callizonus (Mollendorff, 1900) (fig. 93a-b), whose dras- tically narrowed shell aperture and large apertural barriers (fig. 92a) closely approach the typical en- dodontid condition. Even though species of Palline (figs. 94b, e, 95d) have large apertural barriers and somewhat narrowed apertures, the thick terminal genitalia (fig. 93f, h) contrasts with that of Jokajdon and is in the typical charopid pattern. It is quite probable that the variety of genital structures seen in the Charopidae as opposed to the Endodontidae are in large part the result of simple space availability for experimentation. CHAROPID-ENDODONTIDCONCHOLOGICAL COMPARISONS Despite almost complete overlap with regard to most conchological characteristics and their close simi- larity when viewed with the naked eye, there are a few clear differences between the Endodontidae and Charopidae in meristic and structural features. Data in Tables I-III and X and Figures 15-23 summarize both similarities and differences in some basic para- meters. Figures 15-22 were prepared several years ago and omit data on the Lau Archipelago endodontids, Priceconcha tuuuthaensis Solem (1973d) and Thauma- todon spirrhymatum Solem (1973d). Their inclusion would not have changed the results significantly. In addition, four taxa of endodontids were not seen or measured. These two exceptions account for the difference between the 185 taxa listed by Solem (1976b, p. 9, table IV) and the 179 listed as measured in this tabular comparison. For ease in visual comparison, in Figures 15-22 the actual numbers of species for each graph unit have been converted into percentages of the total measured within that family so that the graphs will be directly comparable. Otherwise the difference between 179 en- dodontids and 95 charopids would make visual com- parisons difficult. Table X shows the median mean value and total range of mean values for the species measured within each family. The distinctly larger size and higher whorl count of the Endodontidae is evident, whereas their near identity in H/D ratios, D/U ratios, and mean rib counts is surprising. The greater ribs/mm, in the Charopidae directly correlate with their smaller size. TABLE X. - CONCHOLOGICAL COMPARISONS OF ENDODONTIDAE AND CHAROPIDAE Number of species level taxa measured Median mean height (range) Median mean diameter (range) Median mean whorl count (range) Median mean H/D ratio (range) Median mean D/U ratio (range) Median mean ribs on body whorl (range) Median mean ribs/mm. Taxa with ribs reduced PACIFIC ISLANDS. Endodontidae 179 1.48(0.92-7.26) 3.77(1.7-12.3) 5 1/2+ (3 5/8-8 1/8) 0.531 (0.344-0.789) 3.94 (1.68-closed) 91.0(19-202) 9.2(1.4-40) 39(21.8%) Charopidae 95 1.51(0.48-3.69) 2.76(1.07-7.52) 4 1/8- (3 1/8-6 5/8) 0.523 (0.365-0.801) 3.94 (2.03-closed) 91.2(19-225) 11.3(1.7-37) 10(10.5%) The greater degree of rib reduction in the Endodon- tidae and the reduction in percentage of taxa with apertural barriers in the Charopidae also indicate major differences. More detailed comments can be made from the data in Figures 15-23. Mean height distribution (fig. 15) shows a slight Endodontidae offset for most species, then an extended high-spired portion that far sur- Charopidae Endodontidae 175 275 375 Mean Height 4.75 575 over 65 38 FIG. 15. Mean shell height distribution in the Endodontidae and Charopidae. passes the maximum height recorded for the Char- opidae. The three highest spired Charopidae are Lauopa mbalauuana, Tuimalila infundibulus (Hom- bron & Jacquinot, 1841), and T. pilsbryi. All three have average to only slightly increased whorl counts, and only the latter has a high SP/BWW ratio (mean 2.23). Their large height is the result of size increase alone, rather than change in shell form. In contrast, the high-shelled Endodontidae mostly are species with umbilical brood chambers and increased whorl counts, plus the few Nesodiscus and pre-brood chamber En- dodonta (Solem, 1976b, pp. 27-30). Mean diameter distribution (fig. 16) shows a greater similarity between the two families, both hav- ing a significant right extension of the frequency curve. The sharper peak for the Charopidae relates to CHAROPID-ENDODONTID CONCHOLOGICAL COMPARISONS 39 30 25 - 20 - Charopidae Endodontidae 1.75 2.75 3.75 475 Mean Diameter 5-75 6.75 7.75 over 8.0 FIG. 16. Mean shell diameter distribution in the Endodontidae and Charopidae. the speciose genus Sinployea, which accounts for more than half of the total taxa reviewed. The secondary peak around 4.75 mm. results from some of the large Rarotongan and Samoan Sinployea, plus Himero- concha, Russatus, and large Semperdon from Micro- nesia. Both families have a few taxa that exhibit gigantism compared with the average species. Mean H/D ratio distribution (fig. 17) differs only because of the brood chamber taxa in the Endodontidae (Solem, 1976b, p. 29, fig. 19), producing a slight bulge 30 25 20 ~ I o 15 - I -Charopidae Endodontidae 0.325 0.425 0.525 0-625 Mean H/D Ratio 0.725 0.825 FIG. 17. Mean height/diameter ratio distribution in the En- dodontidae and Charopidae. in the higher ratios for that family. The only very high-spired charopids are the nearly scalariform Ba humbugi (fig. 74b), Ladronellum mariannarum (Qua- dras & Mollendorff, 1894) (fig. lOlb), and Semperdon kororensis (Beddome, 1889) (fig. 98b). Only Ladronel- lum has a normally open umbilicus (fig. lOlc), whereas the others have a closed umbilicus that normally in- creases the H/D ratio dramatically (Solem, 1976b, p. 25, fig. 15). The most dramatic difference is shown by the mean whorl count distribution (fig. 18). The only charopids to exceed a mean whorl count of five whorls Charopidae Endodontidae ( I 10 3 1 '. 4'i 5'', 6'* Mean Whorls FIG. 18. Mean whorl count distribution in the Endodontidae and Charopidae. are the Micronesian Semperdon kororensis (Beddome, 1889) (fig. 98a) with 5V& average, and the Rarotongan Sinployea planospira (Garrett, 1881) (fig. 46d) with 6% whorl average. Although there is considerable overlap in mean whorl counts of five whorls or less, the drama- tic increase in higher mean whorl counts in the En- dodontidae is quite clear. A fair portion of this increase to the right correlates with brood chamber formation (Solem, 1976b, p. 29, fig. 20). Means of D/U ratio for the Endodontidae in Figure 19 exclude the brood chamber taxa entirely, hence the two families show very similar patterns. Because egg laying in the umbilicus is a family characteristic in the 40 SOLEM: ENDODONTOID LAND SNAILS n 25- 20- Charopidae Endodonlidae 225 325 4.25 525 6.25 7- 11- 21- crack cloud 10 20 50 X D/U Ratio FIG. 19. Mean diameter/umbilical width ratio distribution in the Endodontidae and Charopidae. Endodontidae (Solem, 1976b, pp. 100-101), extreme narrowing or closure of the umbilicus should be a rarer event in the Endodontidae than in the Charopidae. Surprisingly enough, it is not. The only Charopidae with a closed or nearly closed umbilicus are Semperdon kororensis (Beddome, 1889) (fig. 98c) from Palau, Sinployea clista (fig. 51c) from Samoa, and Be humbugi (fig. 74c) from Viti Levu, Fiji. Sinployea clausa (fig. 5 If) from Samoa has an extremely narrowed um- bilicus. Most other species have the umbilicus much more widely opened, and only in the Lau Archipelago Sinployea adposita (Mousson, 1870) (fig. 69f) andRuss- atus nigrescens (Mollendorff, 1900) (fig. 89f) from Ponape is the umbilicus very narrow. In the Endodon- tidae (Solem, 1976b, pp. 26, 491-492) there are 10 taxa with closed and 10 with barely perforate umbilici. Most of these are found on Rapa and Mangareva, areas far from the main areas of distribution and possibly from the natural occurrence of potential egg predators that occupy the leaf litter. Hence, the unexpected high occurrence of closed or barely perforate umbilici in the Endodontidae is a geographic phenomenon. The degree of spire protrusion, as measured by the SP/BWW ratio, is quite different for the two families (fig. 20). The Endodontidae are in general much higher spired. This relates primarily to the taxa reaching the Nesodiscus and brood chamber levels of organization. The most elevated Charopidae are Ba humbugi from Fiji (fig. 74b), Ladronellum mariannarum (Quadras & Mollendorff, 1894) from Guam (fig. lOlb), Sinployea angularis from Fiji (fig. 64e), and Vatusila nayauana from Fiji (fig. 82e). Because so few Pacific Island Charopidae have elevated spires, the type of analysis done for the Endodontidae (Solem, 1976b, p. 25, fig. 15) in which spire protrusion was correlated with varia- Cfiaropidae -.- Endodontidae . depressed Hal 0.05 0.15 0.25 035 04 055 065 075 085 0.95 overlO Spire height/body whorl width ratio FIG. 20. Mean spire height/body whorl width ratio distribution in the Endodontidae and Charopidae. tions in shell height, diameter, H/D ratio, and D/U ratio is not presented since the differences were so slight. Comparatively few Charopidae have a flat or clearly depressed spire. Only Roimontis tolotomensis has a clearly sunken spire. The striking similarity in median mean rib counts on the body whorl (table X) and the very similar distri- bution of mean rib counts in the two families (fig. 21) was not anticipated. It does suggest that the hypoth- esized function of the ribbing to reduce adherence of particles to the shell surface (Solem, 1976b, p. 50) may be correct. Rib reduction in the Endodontidae is primarily size correlated. Species with a mean diame- ter of 4.75 mm. or more frequently show a marked de- gree of rib reduction (Solem, 1976b, pp. 46-50, tables Cliaiofxd* Endodontidae 10 - 190 over reduce) 200 70 110 ISO Ribs on Body Whorl FIG. 21. Mean ribs on body whorl distribution in the Endodon tidae and Charopidae. CHAROPID-ENDODONTID CONCHOLOGICAL COMPARISONS 41 XVII- XIX). There are comparatively few taxa in the Charopidae that show major rib reduction. The 10 are not phyletically correlated and show reduction in different degrees and ways. In Himeroconcha rotula (Quadras & Mollendorff, 1894) and H. lamlanensis from Guam, plus Sinployea rudis (Garrett, 1872) and S. harveyensis (Garrett, 1872) from Rarotonga, the sculpture is reduced to irregularity on the body whorl, with the major ribs becoming too crowded and too ir- regular to count well before the lip. In Sinployea ir- regularis (Garrett, 1887) and Ba humbugi from Fiji plus Russatus nigrescens (Mollendorff, 1900) from Ponape, the postapical sculpture is highly irregular, and occasionally a rib with a high lamellar extension appears. In Sinployea recursa from Fiji and Himeroconcha quadrasi (Mollendorff, 1894) and H. fusca (Quadras & Mollendorff, 1894) from Guam, the sculpture is reduced in prominence on both the spire and body whorl and rapidly becomes indistinguishable from growth lines after an initial portion where the ribs are large enough to be counted. The four Himeroconcha do show a pattern of size-associated re- duction in rib prominence, since the two smaller species have more prominent spire sculpture (fig. 104a-f) than do the two larger taxa (fig. 105a-f). There is no data available concerning the ecology of Himeroconcha, so that the reasons for this pattern of sculpture reduction are unknown. Russatus, Ba, and Sinployea irregularis (Garrett, 1887) are known to be terrestrial in habitat, and Garrett (1872, pp. 227-228) reported that both S. rudis (Garrett, 1872) and S. har- veyensis (Garrett, 1872) were collected under rotting wood. Only dead examples of the Lau Archipelago Sinployea recursa are known, although I have specu- lated that it could be an arboreal species. The overall pattern of rib reduction in the Charopidae is less clearly size linked than in the En- dodontidae. Three of the 10 species, S. recursa, S. ir- regularis (Garrett, 1887), and Ba humbugi are within 0.11 mm. of the median mean diameter, the other seven (38.9%) are among the 18 over 3.75 mm. in diameter, and three (37.5%) are among the eight species that average more than 4.75 mm. in diameter. They are Russatus nigrescens (Mollendorff, 1900), Himeroconcha fusca (Quadras & Mollendorff, 1894), and H. quadrasi (Mollendorff, 1894). In the Endodon- tidae, 50% of the taxa averaging over 4.76 mm. in diameter have reduced sculpture (Solem, 1976b, p. 47, table XVIII). The variation in mean ribs/mm, on the body whorl (fig. 22) is equivalent in the two families, once allow- ance is made for the smaller mean size of the Charo- pidae. As in the Endodontidae (Solem, 1976b, pp. 44-45, tables XV- XVI), the smaller the mean diame- ter, in general the more numerous are the radial ribs (see table X). In summary, basic shell size, shape, and radial sculpture spacing and frequency in both families are more similar than dissimilar, once allowance is made Charopidae [ndodonlidae J I H 15 Ribs/mm I I over reduced 22 FIG. 22. Mean ribs/mm, on the body whorl distribution in the Endodontidae and Charopidae. for the changes in the Endodontidae caused by the for- mation of a brood chamber. The Endodontidae are larger and with a higher whorl count than the Charopidae, but in basic form and character of the ra- dial sculpture they are very similar. This is not sur- prising because both are basically found in litter, under stones, and in or under rotting wood. A much higher number of Charopidae than Endodontidae has been taken in semiarboreal to arboreal situations. In addition to gross size, shape, and sculpture comparisons, an important question is whether there is a different pattern of growth between the two families. With the complications introduced by variations in spire protrusion, body whorl descension, umbilical width, whorl counts, and sculptural protrusions, it is difficult to find a simple index of whorl increment pat- tern that will at least indicate the pattern of whorl width increase. Figure 23 attempts this through use of a crude measure of size increment plotted against whorl count of the figured individual. To try and minimize difficulties in interpretation, brood chamber taxa and the very high-spired members ofAaadonta in the Endodontidae plus the high-spired Ba humbugi in the Charopidae have been omitted. Inclusion of these would have added data points to the lower right of the figure. Because the increase in shell diameter is less when the spire is elevated, the results would have been artificially skewed toward the right. The basic data for Figure 23 were taken from the top views of shells as published in Solem (1976b) and this report. Because top views were not published for species of Anceyodonta, nearly all Cookeconcha, Opan- ara, Rhysoconcha, Ruatara, Orangia, Taipidon, Planudonta, Rikitea, and Nesophila, this is only a par- tial sampling of the Endodontidae. In contrast, nearly all of the Charopidae have been figured in top view. On each illustration, two measurements were made: the width of the first whorl from suture to suture and then the diameter of the entire shell. In addition, the 42 SOLEM: ENDODONTOID LAND SNAILS 27- 26 25- 24 23 22 21 20 19 18 I 16 Endodontidae Charopidae D D D O D D DD DO D D DD Q OOD 6 D 5 2 D D D D 3 n D DO OOD dS 13 12 11 D DQ D rm 4 D I I OT D DO DQ D DD D D 10 D an Whorl count FIG. 23. Patterns of whorl width increment in selected Endodontidae and Charopidae. Brood chamber, very high spired, and unillustrated top views of species in the Endodontidae are omitted, as is the high-spired Ba humbugi in the Charopidae. Unusual taxa are: (1) Minidonta manuaensis Solem, 1976; (2) Mautodontha aoraiensis Solem, 1976; (3) Australdonta pseudplanulata Solem, 1976; (4) A. pharcata Solem, 1976; (5) Cookeconcha stellulus (Gould, 1844); (6) Thaumatodon laddi Solem, 1976; (7) Zyzzyxdonta alata Solem, 1976; (8) Aaadonta kinlochi Solem, 1976; (9)Sinployea canalis (Garrett, 1872); (10) S . planospira (Garrett, 1881); (11) Russatus nigrescens (Mollendorff, 1900); (12) Semperdon kororensis (Beddome, 1889); and (13) Minidonta micraconica Solem, 1976. number of whorls in the illustrated specimen was re- corded. The width of the first whorl was divided by the total diameter to give a percentage of the total diame- ter taken up by the initial nuclear whorl. This is, of course, partly inaccurate, because the suture-to-suture distance will be less than the periphery-to-periphery distance. The error will be approximately the same for each species, so that the degree of error will be in the same direction. Use of a percentage index of total diameter plotted against the actual whorl count gives an approximate indication of basic whorl width incre- ments. It is evident that members of the Endodontidae have a tighter pattern of coiling at the same whorl count than do the Charopidae. The higher the percent- age of the first nuclear whorl width, the slower is the rate of whorl width increment for succeeding whorls. There is, in general, a rather clear separation between the two family units. The exceptions from the general pattern in the Charopidae are a few species with un- usual coiling patterns: (number 9 on fig. 23) Sinployea canalis (Garrett, 1872) (fig. 49a-c), which has a later- ally compressed whorl profile and increased whorl count; (number 10) S. planospira (Garrett, 1881) (fig. CHAROPID-ENDODONTID CONCHOLOGICAL COMPARISONS 43 46d-h), which is laterally compressed, has increased whorl count, and very tight coiling; and (number 12) Semperdon kororensis (Beddome, 1889) (fig. 98a-c), with its raised spire, lateral compression, and closed umbilicus. At the other extreme, (number 11) Russatus nigrescens (Mollendorff, 1900) (fig. 89d-f), which has a flat spire, reduced whorl count, and reduced sculpture, is offset to the left considerably. The unusual Endodon- tidae include (number 1) Minidonta manuaensis Solem (1976b, p. 131, fig. 62a) and (number 13) Minidonta micraconica Solem (1976b, p. 138, fig. 65a-c) with re- duced whorl counts and abnormally flat-spired taxa in normally more elevated genera such as (number 2) Mautodontha aoraiensis Solem (1976b, p. 160, fig. 74d-f), (number 3) Australdonta pseudplanulata Solem (1976b, p. 295, fig. 127d-f), and (number 4) A.pharcata Solem (1976b, p. 313, fig. 137a-c). Taxa whose diame- ter was increased by greatly enlarged radial ribs in- clude (number 5) Cookeconcha stellulus (Gould, 1844) (Solem, 1976b, p. 218, fig. 93a-c) and (number 7) Zyz- zyxdonta alata Solem (1976b, p. 466, fig. 198a-c). Aaadonta kinlochi Solem (1976b, p. 486, fig. 208a-c, number 8) is a flat-spired species in a genus that is normally highly elevated, whereas (number 6) Thaumatodon laddi Solem (1976b, p. 452, fig. 193d-f) is a flat-spired species with slightly protruded periphery. Thus, all exceptions are readily explained as representing special situations. Correlated with the more rapid whorl width in- crement in the Charopidae is a greater increase in cross-sectional area of the whorl profile. Without sec- tioning shells and measuring the areas of the whorls, quantification of this difference is not feasible. The fact of the greater cross-sectional area and concomitant in- crease in linear wall distance at any given point un- doubtedly has had major effects on the placement of pallial organs in relation to each other and their length. It also has permitted thickening of genital or- gans over the endodontid condition. Discussion of these changes was dealt with under the patterns of anatomi- cal variation. On the submicroscopic level, there are three major differences between the Endodontidae and Charopidae found on the Pacific Islands. The apical sculpture of the shell, the method of forming the postapical shell sculpture, and the way in which the apertural barriers are formed and armed with microdenticles are different. These differences have been reviewed in part elsewhere (Solem, 1969d). Only a brief summary is in- cluded here. In the Endodontidae, the apical sculpture consists of prominent radial ribs, with or without microriblets in between, plus very fine spiral cords that are best termed "squiggly" (Solem, 1976b, pp. 35-41, figs. 25-31). In one genus, Aaadonta Solem (1976b, pp. 38-39, figs. 28-29), the major radial sculpture has been lost, only the microradials are left on the postapi- cal whorls, and only the squiggly spiral cords are left on the apex. In the Pacific Island Charopidae, the basic apical sculpture is of strong spiral cords, typically as seen in Sinployea modicella (Ferussac, 1840) (fig. la-c). They can be reduced in prominence and in- creased in number, as in Sinployea peasei (fig. 2a-b), or combined with a secondary sculpture of low radial swellings, as in Tuimalila pilsbryi (fig. 2c). The only exceptions on the Pacific Islands concern the rotadis- cine genus Microcharopa, in which the spiral apical cords are broken up into short, twisted segments (fig. 4a-e), and the widespread Discocharopa (fig. 5), in which both apical and postapical sculpture consists of fine radial ribs and there is no trace of spiral sculpture. The situation becomes much more complicated in the charopid taxa of Australia and New Zealand where there have been many experiments in shell sculpture (Solem, unpublished data). However, in relation to the Pacific Island taxa, the sharp division into taxa with spiral cords (Charopidae) and taxa with radial ribs (Endodontidae) holds with a few secondary exceptions. Under optical magnification, Aaadonta would be con- fused with the charopid condition and Tuimalila with the endodontid condition, but SEM studies show that these apparent exceptions are secondary modifications. Except for obvious periostracal setae and exten- sions in such taxa as Cookeconcha decussatulus (Pease, 1866) (Solem, 1976b, p. 36, fig. 26a-c; p. 40, fig. 30b), the postapical sculpture in the Endodontidae appar- ently is formed by a thin template of periostracum, with most of the sculpture thickness consisting of un- derlying calcium layers. This holds even for the fine apical features (Solem, 1976b, p. 35, fig. 25d). In con- trast, the microsculpture and much of the major rib projection in the Charopidae consists entirely of perio- stracal materials. Frequently the only calcareous sculptural element will be a swelling underneath each major radial rib. An example of this is seen in Sinployea uicaria vicaria (Mousson, 1871) (fig. 59). The same pattern of structure seems to hold true for at least many of the New Zealand and Australian Charopidae. This makes the similarity in gross sculpture effect be- tween the two families even more remarkable. An initial review of the apertural barrier differences between the Endodontidae and Charopidae was given in Solem (1973b). The following summary is taken from that paper, Solem (1976b, pp. 52-72), and pp. 15-23. In the Endodontidae, the barriers show a uniform pattern of structure, growth, and microden- ticulation. The parietal and columellar walls will have weak pustulations, as is typical of many taxa in a vari- ety of families (Solem, 1972c), but the characteristic feature on the barrier tops is a series of additive tri- angular microdenticulations (Solem, 1973b, figs. 7-13, 23-24; Solem, 1976b, p. 55, fig. 39d-e; p. 64, figs. 40, 41a-c). It apparently does not matter whether the ex- panded upper edge of the barrier is continuous or broken up into a series of expanded beads as in Thaumatodon, Aaadonta, and Zyzzyxdonta (Solem, 1976b, p. 457, fig. 194b-e; p. 466, fig. 198b; p. 475, fig. 203b, e). The pattern of minute triangular denticles 44 SOLEM: ENDODONTOID LAND SNAILS that face toward the aperture and are additions to the surface of the barrier, rather than an outgrowth from it, holds. The only known exceptions are in certain of the very large species of Cookeconcha and Endodonta, where the denticulations on the upper edge of the bar- riers are blunt-tipped (Solem, 1976b, p. 66, fig. 42d-f) although retaining their triangular shape on the sides of the barriers, and in other genera such asNesodiscus, where the barriers are greatly reduced in both size and number. These barriers may be secondarily without denticulations. I have thus concluded that the barriers seen in members of the Endodontidae had a common origin and that reduction or loss of barriers is a secon- dary phenomenon in that family. In the Charopidae (pp. 15-17), in contrast, three of the four major groups with barrier-equipped genera found on the Pacific Islands show barriers that differ in size, position, growth, and microdenticulations (Solem, 1973b, pp. 303-304, figs. 4-6, 14-22). The latter basi- cally are crystals growing out of the barrier surface and are blunt tipped. It is thus concluded that aper- tural barriers have evolved independently in each of these lineages, despite the fact that SEM observations could not be made on the Smp/ojea-derivative genera. Their barriers show such obvious differences from the other three groups (figs. 80-83, 85) that I have no doubt of their independent origin. More detailed dis- cussion of the barrier structure and variation has been given in the Patterns of Morphological Variation (pp. 15-23). In the Endodontidae the great majority of the species have barriers on the palatal wall, with a bladelike shape and gradual anterior descension. In the Charopidae only four species have bladelike palatal barriers, and 13 have purely crescent-shaped barriers. The frequency of barrier occurrence is lower in the Charopidae than in the Endodontidae. In the latter, including taxa not seen but illustrated in older litera- ture, 184 out of 185 (99.5%) have parietal barriers and 163 (88.1%) have palatal barriers. In the Charopidae 32 (33.7%) of 95 taxa have parietals, and 27 (28.4%) have palatal barriers. Thus the families differ signifi- cantly not only in whether or not barriers are present in the aperture but also in how they are formed, the normal shape of the palatals, and the nature of the microdenticulations on the barrier edges. Despite great optical similarity in size, color, shape, presence of apertural barriers in many species, and presence of fine radial ribbing in at least the smaller taxa, differences in whorl count (fig. 18), growth pattern of whorl width (fig. 23), apical sculpture, method of formation for the postapical sculpture, and details of the apertural barrier structure and microdenticulations serve to distinguish the two families on conchological grounds. HABITAT RANGE AND EXTINCTION In contrast with the Endodontidae (Solem, 1976b, pp. 100-101), the Charopidae have been able to exploit the semiarboreal and arboreal habitats as well as re- maining well represented in the ground stratum. Their closed secondary ureter has been hypothesized (Solem, 1976b, p. 100) as the reason the charopids have been able to become successfully semiarboreal. A second fac- tor in their success is the apparent lack of umbilical egg deposition by the charopids. Of all the Pacific Is- land charopid specimens studied, only one example of Tuimalila infundibulus (Hombron & Jacquinot, 1841) had a snail egg capsule inside the umbilicus, whereas this is the routine pattern in the Endodontidae. Of ex- tralimital taxa, the New Zealand charopids, Fectola marsupialis Powell, 1941, and Aeschrodomus worleyi Powell, 1928, have the umbilical egg-laying pattern (see Climo, 1969a, pp. 218-219; 1970, p. 306). On such drastically ecologically altered islands as Viti Levu, Tahiti, and Moorea, specimens ofSinployea and Ba have been collected from lowland vegetational remnants in the 1960s and 1970s. They have not, how- ever, been found where the ground stratum has been severely and regularly disturbed by chickens and/or pigs. In parts of Samoa and on the Lau Archipelago of Fiji the charopids still are common and relatively abundant. Only on Rarotonga is there clear evidence of major charopid extinctions within a century. Two sepa- rate collecting efforts on Rarotonga in the mid-1960s failed to find nine out of 10 Sinployea species collected in the 1860s and 1870s and five out of six endodontids reported from the same island. All of the missing taxa were described as ground stratum inland species. The one collected endodontid occupies the storm line coral boulder habitat, and the one previously described Sinployea is at least occasionally semiarboreal. The actual egg-laying habits of the Pacific Island Charopidae are unknown, but their nondependence upon the umbilicus for egg-laying probably has effec- tively lowered ant predation, which apparently was a major factor in the mass extinction of the Pacific Island Endodontidae. The semiarboreal to arboreal habitat range extension of the Charopidae also has lessened human impact through introductions, because arboreal snails of the Pacific Islands, in general, have been able to adapt to plantations of bananas and even coffee, whereas strictly ground-dwelling taxa are absent from these human-introduced environments. The charopids are abundant in Australia, South Africa, New Zealand, New Caledonia, and parts of South America all areas in which ants are common. I thus expect that the ability of the Pacific Island charopids to survive despite the introduction of preda- cious ants is a carryover from conditions in their areas of greater abundance rather than a new adaptation. Since ants first appeared in the Cretaceous fossil rec- ord (Wilson et al., 1967), and have an extensive fossil record of living tribes and genera from the Oligocene and Miocene, their coexistence with the basic charopid stock seems highly probable. The charopids thus occupy a wider range of habitats than do the Endodontidae and persist in fair variety despite the great ecological alterations of the past century on the Pacific Islands. 45 PHYLOGENY AND CLASSIFICATION Reviews of the hypothesized early evolution and progressive trends among land snails have been given previously (Solem, 1974; 1976b, pp. 102-104; 1978b). The position of the Endodontidae as the most gen- eralized extant group of the Sigmurethra, the absence of any family unit from which the endodontid anatomi- cal structures can be derived, and the positioning of both the Charopidae and Endodontidae within the Arionacea complex have been discussed earlier and will not be rejustified here. The Charopidae are more advanced than the En- dodontidae in having a closed secondary ureter, fused prostate-uterine tubes with a common lumen, a re- duced number of lobes in the ovotestis, and in transfer- ring sperm in a packet (see Ba humbugl, fig. 75f-g), which correlates with the general presence of an epiphallus. Charopids are more complex than the En- dodontidae in having great elaboration of terminal genital structures and in the variety of their radular structures. Recent dissections of some undescribed Western Australian taxa have revealed genera that are partly transitional between the Endodontidae and Charopidae. These will be described and discussed elsewhere. It is premature to discuss in detail the probable phylogenetic relationships among the endodontoid families. After synoptic coverage of the Austro- Zelandic taxa, such a discussion will be profitable. At this time I choose to end the discussion of interfamily phylogeny by emphasizing that the Charopidae are more advanced than and derivable from the Endodont- idae. The phylogenetic position of the Punctidae in re- lation to the other families is still uncertain. 46 FAMILY CLASSIFICATION OF THE ENDODONTOIDS A list of subfamily and family names available for members of this complex was given in Solem (1976b, p. 105). I recognize five families and group the assignable names as below. The name Patulinae Tryon (1866, p. 243) is discarded (see Solem, 1976b, p. 105). The af- finities of the Indian Thysanotinae Godwin-Austen, 1907, are still unknown. Available literature informa- tion on their anatomy is not adequate, and I have not been able to obtain preserved materials for dissection. I have not dissected one of Iredale's family units, the Hedleyoconchidae, but on the basis of shell structures have no hesitation in lumping it with the Charopidae pending further study. Definitions and discussion of the Helicodiscidae (see Solem, 1975) and Discidae (Solem, in preparation A) are not included here, since both families are extralimital to this study. Definitions of both the Punctidae and Charopidae are included below in the systematic review; the Endodontidae were defined by Solem (1976b, p. 121). The proposed synonymization of family level units, in chronological order, is: Family Punctidae (Morse, 1864, p. 27) + Laominae Suter (1913, p. 732) + Paralaomidae Iredale (1941a, p. 263) Family Charopidae Mutton (1884b, p. 199) + Phenacohelicidae Suter (1892a, p. 270) + Otoconchinae Cockerel! (1893, pp. 188, 205) + Flammulinidae Crosse (1894, p. 210) + Rotadiscinae H. B. Baker (1927, pp. 226, 230) + Amphidoxinae Thiele (1931, p. 575) + Dipnelicidae Iredale (1937b, pp. 22-23) + Hedleyoconchidae Iredale (1942, pp. 34-35) + Pseudocharopidae Iredale (1944, p. 312) Family Endodontidae Pilsbry, 1895 (Pilsbry, 1893-1895, p. xxviii) Family Helicodiscidae Pilsbry, 1927 (in Baker, 1927, pp. 226, 230) + Stenopylinae Thiele (1931, p. 569) Family Discidae Thiele ( 1931, p. 578) + Goniodiscinae Wagner (1927, p. 305) Although I do not accept strict nomenclatural priority for names of families and higher level taxa, the above usage comes close to following this precept. I reject both Patulinae Tryon (1866) and Goniodiscinae (Wagner, 1927); the former name because Patula mostly has been associated with members of the land snail family Oreohelicidae, a totally unrelated group, until finally assigned as a synonym of Discus. Goniodiscus Fitzinger, 1833, is now ranked as a sub- genus or synonym of Discus (see discussion by Forcart, 1957), and I consider it inappropriate to use as a family-level name. The morphological gaps between these families are of the same order of magnitude or greater than the gaps separating recognized families of the Limacacea (see Solem, 1976b, p. 107), although the gaps between the family units certainly are not equal in size. Discussion of subfamily and generic units is de- ferred to the family discussions of the Punctidae and Charopidae. 47 PHYLOGENETIC PROCEDURES I am a pragmatic phylogeneticist in the sense of Mayr and have used neither Hennigian cladistics nor phenetic manipulations in this study. The criteria used are those cited by Solem (1976b, p. 108) and the three- tiered approach to evolutionary change outlined in Solem (1978b). These are suited to the peculiarities of molluscan shell growth, with the ontogenetic develop- ment from embryo on visible in old age, the water- dependent ecology of land snails, and their use of ter- minal genitalia for species recognition. The absence of many fixed reference points within the snail's body and the difficulties of making precise and meaningful measurements on the shell combine to make quantification of many features impossible and thus use of the more sophisticated mathematical ma- nipulations of data matrices impractical. Similarly, in the absence of equivalently detailed analyses of most other land snail families and with the uncertainty as to the overall phylogeny of either characters or taxa, use of cladistic methodology is premature. 48 GENERIC CLASSIFICATION The same type of "pigeon-hole" generic units cited for the Endodontidae (Solem, 1976b, pp. 118-119) existed for the Pacific Island Charopidae. If there were apertural barriers present, the species was referred to the form genus Endodonta or to one of its sections, and if there were no such barriers present, it was placed in the New Zealand genus Charopa. Although Pilsbry (1893-1895, p. 21) had pointed out the artificiality of this division, no significant changes were introduced subsequently. Reference of a new species to Flam- mulina (Russatus nigrescens) by Mollendorff (1900) or attempts to use other described genera than Charopa (Solem, 1959a, 1960) did not alter the essential absence of any meaningful classification on the generic level. Of the 21 genera in the Punctidae and Charopidae reviewed below, only two, Punctum and Discocharopa, have available names. The other 19 are named here. The same criteria for generic separation are used here in the Charopidae as were used in classifying the En- dodontidae into genera (see Solem, 1976b, pp. 119- 120). Thus the generic concepts are directly compar- able and the minimum generic gap equally applied. The difference in frequency of monotypic genera in the two families and presence of one particularly speciose genus in the Charopidae was summarized by Solem (1976b, p. 120). This is a real difference in evolutionary patterns and not an artifact of classification. 49 SYSTEMATIC REVIEW Because the two families are so disparate in the number of species reviewed and the complexities of charopid classification require considerable reference to extralimital taxa, the discussion of charopid phylogeny and classification is subtended to the family review. Following are a formal list of the taxa reviewed and geographic keys to the genera. These sections fol- low the format used in Solem (1976b) as do the species and generic accounts. LIST OF THE TAXA Family Punctidae Morse ( 1864, p. 27) Genus Punctum Morse, 1864 Punctum sp. Society Islands: Tahiti, Mt. Aorai Punctum polynesicum, new species Austral Islands: Tubuai, Raivavae Family Charopidae Hutton (1884b, p. 199) Subfamily Rotadiscinae H. B. Baker (1927, pp. 226, 228) Genus Microcharopa, new genus Microcharopa rnimula, new species Fiji Islands: Viti Levu; Lau Ar- chipelago (Munia, Mothe, Wangava, Nayau, Namuka, Yangasa Levu) Subfamily Charopinae Hutton ( 1884b, p. 199) Genus Discocharopa Iredale, 1913 Discocharopa aperta (Mollendorff, 1888) Philippines, Indonesia, New Guinea, Western Australia, Northern Territory, Queensland, Bismarck Archipelago, New Hebrides, Fiji Islands, Samoan Is- lands, Austral Islands (Rurutu), Society Islands (Borabora) Genus Sinployea, new genus Sinployea modicella (Ferussac, 1840) Society Islands: Moorea Sinployea tahitiensis, new species Society Islands: Tahiti Sinployea lame/licosta (Garrett, 1884) Society Islands: Tahiti Sinployea montana, new species Society Islands: Tahiti Sinployea neglecta, new species Society Islands: Huahine Sinployea sp. Society Islands: Borabora Sinployea atiensis (Pease, 1870) Cook Islands: Atiu, Aitutaki Sinployea andrewi, new species Cook Islands: Mangaia Sinployea peasei, new species Cook Islands: Rarotonga Sinployea avanaensis, new species Cook Islands: Rarotonga Sinployea proximo (Garrett, 1872) Cook Islands: Rarotonga Sinployea planospira (Garrett, 1881) Cook Islands: Rarotonga Sinployea rudis (Garrett, 1872) Cook Islands: Rarotonga Sinployea decorticata (Garrett, 1872) Cook Islands: Rarotonga Sinployea harveyensis (Garrett, 1872) Cook Islands: Rarotonga Sinployea youngi (Garrett, 1872) Cook Islands: Rarotonga Sinployea canalis (Garrett, 1872) Cook Islands: Rarotonga Sinployea otareae (Garrett, 1872) Cook Islands: Rarotonga Sinployea tenuicostata (Garrett, 1872) Cook Islands: Rarotonga Sinployea clausa, new species Samoan Islands: Manu'a Group (Tau) Sinployea clista, new species Samoan Islands: Tutuila, Upolu Sinployea aunuuana, new species Samoan Islands: -Aunuu off Tutuila, possibly Manu'a Group (Tau) Sinployea allecta allecta (Cox, 1870) Samoan Islands: Upolu, Savai'i Sinployea allecta tauensis, new subspecies Samoan Islands: Manu'a Group (Tau) Sinployea complementaria (Mousson, 1865) Samoan Islands: Upolu Sinployea intermedia, new species Swains Island Sinployea vicaria vicaria (Mousson, 1871) Hoorn Islands: Futuna; Tonga Islands: Vava'u, Tongatapu, Eua, Ha'apai Group Sinployea vicaria paucicosta, new subspecies Tonga Islands: Vava'u Sinployea rotumana (Smith, 1897) Rotuma Sinployea ellicensis ellicensis, new species and subspecies Ellice Is- lands: Funafuti Sinployea ellicensis nukulaelaeana, new subspecies Ellice Islands: Nukulaelae Sinployea pseudovicaria, new species Ellice Islands: Vaitupu Sinployea kiisaieana, new species Caroline Islands: Kusaie Sinployea sp. Mariana Islands: Saipan (possibly introduced) Sinployea angularis, new species Fiji Islands: Lau Archipelago (Namuka) Sinployea recursa, new species Fiji Islands: Lau Archipelago (Wangava) Sinployea princei (Liardet, 1876) Fiji Islands: Taveuni, Kandavu Sinployea inermis inermis (Mousson, 1870) Fiji Islands: Lau Ar- chipelago (Vanua Mbalavu, Mango, Kimbombo) Sinployea inermis meridionalis, new subspecies Fiji Islands: Lau Archipelago (Yangasa Levu, Navutu-i-Loma, Aiwa) Sinployea inermis lakembana, new subspecies Fiji Islands: Lau Ar- chipelago (Lakemba) Sinployea lauensis, new species Fiji Islands: Lau Archipelago (Nayau, Yangasa Levu, Navutu-i-Loma, Wangava) Sinployea navutuensis, new species Fiji Islands: Lau Archipelago (Navutu-i-Loma, ? Oneata) Sinployea adposita (Mousson, 1870) Fiji Islands: Lau Archipelago (Oneata, Mothe, Munia, Lakemba, Aiwa, Karoni, Nayau, Yangasa Levu) Sinployea irregularis (Garrett, 1887) Fiji Islands: Viti Levu Sinployea godeffroyana, new species Fiji Islands: Viti Levu Sinployea euryomphala (Solem, 1959) New Hebrides: Espiritu Santo, Maewo, Gaua, Vanua Lava Sinployea solornonensis (Clapp, 1923) Solomon Islands: Ugi Island off San Cristobal Sinployea kuntzi (Solem, 1960) Solomon Islands: Florida Island off Tulagi Sinployea nissani (Dell, 1955) Solomon Islands: Nissan (Green Is- land), north of Bougainville Sinployea novopommerana (Rensch, 1937) Bismarck Archipelago: New Britain Sinployea descendens (Rensch, 1937) Bismarck Archipelago: New Britain Genus Ba, new genus Ba humbugi, new species Fiji Islands: Viti Levu Genus Maafu, new genus Maafu thaumasius, new species Fiji Islands: Lau Archipelago (Nayau) 50 SYSTEMATIC REVIEW 51 Genus Lauopa, new genus Lauopo mbalavuana, new species Fiji Islands: Lau Archipelago (Vanua Mbalavu) Genus Tuimalila, new genus Tuimalila infundibulus (Hombron & Jacquinot, 1841) Tonga Is- lands: Vava'u Tuimalila pilsbryi, new species Tonga Islands: Eua Genus Lagivala, new genus Lagivala davidi (Ladd, 1968) Ellice Islands: Funafuti (fossil) Lagivala vivus, new species Fiji Islands: Viti Levu Lagivala minusculus, new species Fiji Islands: Viti Levu Lagivala macroglyphis (Rensch, 1937) Bismarck Archipelago: New Britain Lagivala microglyphis (Rensch, 1937) Bismarck Archipelago: New Britain Lagivala demani (Tapparone-Canefri, 1883) Indonesia (Timor, Ambon, Aru Islands); West Irian (Misool) Genus Vatusila, new genus Vatusila kondoi, new species Fiji Islands: Lau Archipelago (Nayau) Vatusila nayauana, new species Fiji Islands: Lau Archipelago (Nayau) Vatusila eniwetokensis (Ladd, 1958) Marshall Islands: Eniwetok (fossil) Vatusila tongensis, new species Tonga Islands: Eua Vatusila vaitupuensis, new species Ellice Islands: Vaitupu Vatusila niueana, new species Niue Island Genus Graeffedon, new genus Graeffedon graeffei (Mousson, 1869) Samoan Islands: Upolu Graeffedon savaiiensis, new species Samoan Islands: Savai'i Graeffedon pricei, new species Tonga Islands: Tongatapu Subfamily Trukcharopinae, new subfamily Genus Trukcharopa, new genus Trukcharopa trukana, new species Caroline Islands: Truk, Lukunor Genus Kubaryiellus, new genus Kubaryiellus kubaryi (Mollendorff, 1900) Caroline Islands: Ponape Genus Russatus, new genus Russatus nigrescens (Mollendorff, 1900) Caroline Islands: Ponape Genus Roimontis, new genus Roimontis tolotomensis, new species Caroline Islands: Ponape Genus Palikirus, new genus Palikirus cosmetus, new species Caroline Islands: Ponape Palikirus ponapicus (Mollendorff, 1900) Caroline Islands: Ponape Genus Jokajdon, new genus Jokajdon tumidulus (Mollendorff, 1900) Caroline Islands: Ponape Jokajdon callizonus (Mollendorff, 1900) Caroline Islands: Ponape Genus Palline, new genus Palline notera notera, new species and subspecies Palau Islands: Peleliu Palline notera palauana, new subspecies Palau Islands: Koror Palline notera gianda, new subspecies Palau Islands: Babelthuap Palline biakensis, new species West Irian: Biak Palline micramyla, new species Caroline Islands: Ponape Subfamily Semperdoninae, new subfamily Genus Semperdon, new genus Semperdon uncatus, new species Palau Islands: Angaur, Aulup- tagel, Koror Semperdon xyleborus, new species Palau Islands: Angaur, Koror, Ngemelis, Peleliu Semperdon kororensis (Beddome, 1889) Palau Islands: Koror Semperdon rotanus, new species Mariana Islands: Rota, northern tip of Guam Semperdon heptaptychius (Quadras & Mollendorff, 1894) Mariana Islands: Guam Genus Ladronellum, new genus Ladronellum mariannarum (Quadras & Mollendorff, 1894) Mariana Islands: Guam Genus Himeroconcha, new genus Himeroconcha lamlanentiis, new species Mariana Islands: Guam Himeroconcha rotula (Quadras & Mollendorff, 1894) Mariana Is- lands: Guam Himeroconcha quadrasi (Mollendorff, 1894) Mariana Islands: Guam Himeroconcha fusca (Quadras & Mollendorff, 1894) Mariana Is- lands: Guam Incertae sedis Helix filiola Ferussac, 1840 Tonga Islands Helix oceanica Le Guillou, 1842 Society Islands: Tahiti Helix minutialis Deshayes, 1851 Society Islands: Tahiti Helix multispirata Hombron & Jacquinot, 1852 Tonga Islands: Vava'u Helix rotula Hombron & Jacquinot, 1852 (not Lowe, 1831) Gambier Islands: Mangareva Pithys verecunda Pease, 1870 Marquesas GEOGRAPHIC KEYS TO THE GENERA The following artificial keys supplement those presented for the Endodontidae in Solem (1976b, pp. 124-126). They are designed to enable identification of adult shells. Endodontid taxa are cross-referenced below but usually are not keyed down to the generic level in order to save space and printing costs. Because there are no charopids or punctids known from the Mangareva, Marquesas, Rapa, or Tuamotu Islands, these areas are not included in this set of keys; nor are keys presented for extralimital areas. Additions to this set of keys are the Caroline, Mariana, and main Fijian Islands. For islands not included in the main groups, reference to the list of "odd island" taxa at the end of the keys may help. AUSTRAL ISLANDS 1. Apical whorls with radial ribs 2 Apical whorls with spiral cords Punctum polynesicum (p. 58) 2. Shell minute, diameter less than 1.8 mm.; at most 1 deeply recessed parietal barrier Discocharopa aperta (Mollendorff, 1888) (p. 76) Shell normally much more than 1.8 mm. in diameter; usually several apertural barriers Endodontidae (Solem, 1976b, p. 125). CAROLINE ISLANDS 1. Aperture without barriers 2 Aperture with 1 or more barriers 6 2. Whorls increasing rapidly in width (fig. 89d); mean D/U ratio about 10; mean whorl count about 3 l /e Russatus nigrescens (Mollendorff, 1900) (p. 215) Whorls increasing less rapidly in width (fig. 89a); mean D/U ratio 3.40-4.00; mean whorl count usually more than 3'/2. ...3 52 SOLEM: ENDODONTOID LAND SNAILS 3. TrukorKusaie 4 Ponape 5 4. Kusaie; spire elevated (fig. 63a); mean rib count on body whorl about 91 Sinployea kusaieana (p. 143) Truk; spire barely elevated (fig. 63d); mean rib count on body whorl about 150 Trukcharopa trukana (p. 208) 5. Spire fiat or sunken; ribs on body whorl about 132; mean diame- ter about 3.75 mm Kubaryiellus kubaryl (Mbllendorff, 1900) (p. 212) Spire slightly elevated; ribs on body whorl about 46; mean diameter about 2.2 mm Palikirus ponapicus (Mollendorff, 1900) (p. 221) 6. No columellar or palatal barriers 7 Palatal and/or columellar barriers present 8 7. Only 1 parietal barrier; spire elevated; 25 large ribs on body whorl Palikirus cosmetus (p. 219) 8. Parietal barriers 2; spire sunken; about 60 low ribs on body whorl Jtoimontis tolotomensis (p. 217) Parietal and palatal barriers many and large 9 Barriers consist of 1 low parietal, 2 low palatals Palling micramyla (p. 234) 9. Spire more elevated; ribs larger; body whorl rounded (fig. 92e) Jokajdon tumidulus (Mollendorff, 1900) (p. 223) Spire barely elevated; ribs smaller; body whorl laterally com- pressed behind aperture (fig. 92b) Jokajdon callizonus (Mollendorff, 1900) (p. 226) COOK ISLANDS 1. Apical whorls with radial ribs; at least 1 apertural barrier Endodontidae (Solem, 1976b, p. 125) Apical whorls with spiral cords; no apertural barriers Sinployea, key on p. 98 ELLICE ISLANDS 1. Apical whorls with spiral cords 2 Apical whorls with radial ribs Endodontidae, Thaumatodon decemplicata (Mousson, 1873) (Solem, 1976b, p. 451) 2. Aperture with barriers 3 Aperture without barriers Sinployea, see p. 134 3. Vaitupu; living; 3 small barriers Vatusila vaitupuensis (p. 196) Funafuti; fossil; 6 large barriers Lagivala davidi (Ladd, 1968) (p. 184) FIJI, MAIN ISLANDS 1. Apical whorls with spiral cords 2 Apical whorls with radial ribs Discocharopa aperta (Mollendorff, 1888) (p. 76) 2. Aperture without barriers 3 Aperture with barriers 5 3. Diameter less than 1.3 mm Microcharopa mimula (p. 71) Diameter more than 2.0 mm 4 4. Spire greatly elevated (fig. 74b); umbilicus closed Ba humbugi (p. 172) Spire nearly flat or slightly elevated; umbilicus open Sinployea, see key on p. 147 5. Parietal barriers 3; D/U ratio about 2.80; ribs on body whorl about 62 Lagivala uiuus (p. 185) Parietal barrier 1; D/U ratio about 5.25; ribs on body whorl about 120 Lagivala minusculus (p. 185) HAWAIIAN ISLANDS Apical whorls with radial ribs, or, in larger species, smooth Endodontidae (Solem, 1976b, p. 125) Apical whorls with spiral cords; shell minute Punctum horneri (Ancey, 1904) (p. 57) LAU ARCHIPELAGO, FIJI 1. Apical whorls with radial ribs 2 Apical whorls with spiral cords 3 2. Several apertural barriers; diameter more than 2.5 mm Endodontidae (Solem, 1976b, p. 125) One or no apertural barriers; diameter less than 2 mm Discocharopa aperta (Mollendorff, 1888) (p. 76) 3. Aperture without barriers 4 Aperture with 1 or more barriers 6 4. Diameter more than 2.0 mm 5 Diameter less than 1.3 mm Microcharopa mimula (p. 71) 5. Periphery protruded into a threadlike keel (fig. 76b); ribs on body whorl about 44 and very large Maafu thaumasius (p. 176) Periphery at most angulated (fig. 68e); ribs reduced in size and usually more numerous Sinployea, see key on p. 147 6. Diameter less than 2.0 mm 7 Diameter more than 7.0 mm.; 1 parietal barrier Lauopa mbalavuana (p. 177) 7. Palatal barriers 3; parietal barrier bifid Vatusila kondoi (p. 192) Palatal barrier 1; parietal barrier simple Vatusila nayauana (p. 194) MARIANA ISLANDS 1. Aperture without barriers 2 Aperture with barriers 3 2. Shell small, diameter less than 2.5 mm.; Saipan Sinployea sp. (p. 145) Shell large, 3.8-7.2 mm.; Guam Jfimeroconcha (p. 259) 3. Aperture with 3 very small barriers (fig. lOlb) Ladronellum mariannarum (Quadras & Mbllendorff, 1894) (p. 255) Aperture with 6 to many barriers (fig. 99b, d-e) Semperdon rotanus and S. heptaptychius (Quadras & Mollen- dorff, 1894) (pp. 245, 247) MARSHALL ISLANDS 1. Apical whorls with spiral cords Vatusila eniwetokensis (Ladd, 1958) (p. 195) 2. Apical whorls with radial ribs Endodontidae (Solem, 1976b, p. 125) PALAU ISLANDS 1. Postapical whorls with microradial riblets only, no major radial ribs Endodontidae, Aaadonta (Solem, 1976b, p. 473) Postapical whorls with major ribs and microsculpture 2 2. Shell diameter less than 2.6 mm.; mean whorl count less than 4V* 3 Shell diameter more than 2.7 mm.; mean whorl count more than4V2 5 3. Palatal barriers 4, low ridges (figs. 94e, 95a); Koror and Babel- thuap 4 Palatal barriers 5, high blades (fig. 94b); Peleliu Palline notera notera (p. 229) 4. Palatal barriers 2; Koror Palline notera palauana (p. 231) Palatal barriers 3; Babelthuap. ..Palline notera gianda (p. 231) 5. Palatal barriers many (fig. 98b, e) 6 Palatal barriers 0-1 (fig. 97b) Semperdon uncatus (p. 238) 6. Mean diameter about 4.6 mm.; umbilicus a very narrow slit Semperdon kororensis (Beddome, 1889) (p. 244) Mean diameter 2.9 mm.; umbilicus widely open Semperdon xyleborus (p. 24P SYSTEMATIC REVIEW 53 TABLE XI. - SPECIES NUMBERS OF ENDODONTOID LAND SNAILS FROM SELECTED GEOGRAPHIC AREAS PUNCTIDAE CHAROPIDAE Polynesia, Fiji Hawaii 1 Known 1 1 Melanesia New Guinea Indonesia New Caledonia'* Lord Howe Id . Norfolk Id. Kermadec Is. 8 Australia Tasmania Victoria N. S. Wales Queensland S. Australia N. Territory 7 2 80 14 9 3 4 1 6 St. Helena Neotropica and Juan Fernandez New Known 29 16 11 9 49 19 8 7 147 38 18 28 14 2 3 10 114 11 55(?) 2 TOTALS 646 Solem (1976b, this monograph, in preparation) Solem (1959a, 1960, 1962, 1963, this monograph) 3 Solem (1958b, I970a) 4 Solem (1961) Collecting by L. Price and P. Bouchet Iredale (1944) and subsequent collecting Iredale (1945) and subsequent collecting 8 Iredale (1913) 9 Powell (1976) Iredale (1937a, 1937c, 1939) U Burch (1976) Collecting by A. Solem New ENDODONTIDAE Known New 143 31 11 50(7, 5 ' 12 ?? 14 3 290 293 Connolly (1939); Solem (1970c) Solem (1978a) Various papers and Odhner (1922) Pilsbry (1948) and various papers F. Clirao, personal communication SAMOAN ISLANDS 1. Apical whorls with radial ribs 2 Apical whorls with spiral cords 3 2. Aperture without palatal barriers; umbilicus very wide Discocharopa aperta (Mollendorff, 1888) (p. 76) Aperture with palatal barriers; umbilicus rather narrow Endodontidae (Solem, 1976b, p. 126) 3. Aperture with prominent barriers 4 Aperture without barriers Sinployea, see key on p. 119 4. Shell diameter about 3.3 mm.; whorls 4Vi; Savai'i Graeffedon savaiiensis (p. 202) Shell diameter over 5.0 mm.; whorls about 5; Upolu Graeffedon graeffei (Mousson, 1869) (p. 200) SOCIETY ISLANDS 1. Apical whorls with radial ribs 2 Apical whorls with spiral cords 3 2. Shell diameter less than 2.0 mm.; at most 1 deeply recessed parietal barrier Discocharopa aperta (Mollendorff, 1888) (p. 76) Shell diameter over 2.25 mm.; usually several parietal barriers or shell diameter over 3.0 mm Endodontidae (Solem, 1976b, p. 126) 3. Shell minute, diameter less than 1.5 mm....Punctum sp. (p. 57) Shell larger, diameter over 2.4 mm Sinployea, see p. 86 TONGA 1. Apical whorls with spiral cords predominating, at most low ra- dial swellings 2 Apical whorls with narrow radial ribs Endodontidae (Solem, 1976b, p. 126) 2. Aperture without barriers 3 Aperture with prominent barriers 5 3. Diameter less than 3.5 mm Sinployea, see key on p. 134 Diameter more than 5.0 mm 4 4. Eua; mean diameter about 6.7 mm. ..Tuimalila pilsbryi (p. 179) Vavau; mean diameter about 5.7 mm Tuimalila infundibulus (Hombron & Jacquinot, 1841) (p. 181) 5. Palatal wall bearing a broad ridge (fig. 83b); Eua Vatusila tongensis (p. 196) Palatal wall with 3 barriers (fig. 88e); Tongatapu Graeffedon pricei (p. 205) A few species are known from islands that do not fall into the major groups. These species are: Futuna Sinployea vicaria vicaria (Mousson, 1871) (p. 134) Niue Vatusila niueana (p. 198) Rotuma Sinployea rotumana (Smith, 1897) (p. 138) Swains Sinployea intermedia (p. 131) It must be emphasized that the above keys are adequate for identification of adult specimens in which the sculpture and apertural barriers can be observed. They are less adequate for juveniles and quite in- adequate for very worn examples. It still will be possi- ble to collect new species in areas such as the Caroline, Palau, Mariana, Lau, and main Fijian Islands that may or may not key out to a correct genus. FAMILY PUNCTIDAE Generally minute to small endodontoids, primitively with spiral apical sculpture and a combination of macro- and microradial post- apical sculpture that is periostracal in origin. Sculpture altered in larger taxa. Whorl counts and shape highly variable, particularly in New Zealand taxa. Umbilicus open, regularly decoiling to nearly closed. Color monochrome to highly tessellated. A few taxa develop apertural barriers, but most lack barriers. Foot of animal undivided. Prominant pedal and suprapedal grooves unite above tail without forming a caudal foss. Radula with small tricuspid central. 54 SOLEM: ENDODONTOID LAND SNAILS lateromarginals bicuspid with three minute accessory cusps, major cusps becoming reduced in prominence on outer teeth, but not tend- ing to split into additional cusps. Basal plates long and slender, cusps tending to point directly forward. Jaw of separated, square plates, partly fused in larger species. Pallial complex with a bilobed kidney, lobes generally equal in length, primary ureter originating at apex of pericardia! kidney lobe, reflexing posteriorly as secondary ureter, opening at ureteric pore alongside to moderately behind the anus and just inside pneumostome. Mantle collar often without lobes or lap>. but a glandular extension onto pallial roof present in some taxa. Genital system variable, but apical sections typically as follows: ovotestis of two closely appressed, variously subdivided, teardrop- shaped lobes lying just above stomach apex; hermaphroditic duct a simple, uncoiled tube entering laterally into carrefour swelling; talon with globular head on a short stalk entering apex of carrefour; albumen gland elongate-ovate, deeply indented by loops of intestine and head of spermatheca. Prostate-uterus partly fused to fused, free oviduct short to long, uterus with different glandular zones. Sper- matheca with lower shaft expanded, upper shaft more slender after tapering, head ovate, expanded, lying imbedded in base of albumen gland above pallial cavity apex. Vas deferens a slender tube entering penis complex near insertion of penial retractor muscle to well below muscle insertion. Penis complex with a penis-derived epiphallic sec- tion in many taxa, interior of penis with long and irregular pilasters, often with a short vergic papilla. Epiphallic section, when present, with slightly different textured pilasters. Entire complex slender and tuhelike. Penial retractor muscle arising from diaphragm, inserting on head of penis or epiphallic section, without complexities of inser- tion. Interior of free oviduct and base of spermathecal shaft with irregular folds in the few examined. Digestive and free muscle sys- tems agreeing with endodontid and charopid patterns. No data are recorded on the nervous system. The family unit Punctidae is used here in a re- stricted sense, not as the umbrella for the endodontoids as suggested by Climo (1969a, 1971a, 1978). Pending anatomical confirmation, I include the Punctidae or Punctinae of Northern Hemisphere workers, Laominae in the sense of Suter (1913) and Gabriel (1930), and Paralaomidae of Iredale (1941a, 1944). As a rough es- timate, about 159 named taxa from the southern parts of Australia, New Zealand, subantarctic islands, Holarctic, and scattered African localities would be in- cluded. Only a very few taxa have had other than jaw and radula data recorded in the literature, and it is quite probable that some of the included species and genera will prove to be charopids and some taxa as- signed to the charopids will prove to be punctids. Par- ticularly when shell sculpture becomes reduced, a common pattern in the Austro-Zelandic taxa, assign- ment to a family group on shell features alone becomes extremely hazardous. The anatomy o{Punctum Morse (1864) was illus- trated and discussed by Baker (1927, pp. 227-228, pi. 16, figs. 8-12) then reprinted in Pilsbry (1948, pp. 641-643, fig. 349a-f). Data on the shell sculpture for- mation was given by Solem (1977b, pp. 150-152, figs. 4-6, 11-12). Anatomical data are set forth here on Punctum (fig. 24a-b), the New Zealand Laoma leimonias (Gray, 1850) (fig. 25a-h), Western Austra- lian Paralaoma aprica Iredale, 1939 (fig. 26a-c), and Tasmanian rr Paralaorna" coesa (Legrand, 1871) (fig. 27a-d). These extralimital taxa are illustrated to show the essential unity of structure within the Punctidae and to provide data for comparing charopid patterns. FIG. 24. Partial anatomy ofPunctum minutissimum (Lea). Cedar bog, Woodburn Road, 4 miles southwest of Urbana, Champaign County, Ohio. E. Keferl! X-20-1969. FMNH 151102: a, pallial region; b, apical genitalia with albumen gland removed. No scale lines pre- pared. (CW). To my knowledge, no other taxa have been illustrated in equivalent detail. An anatomical description of Laoma leimonias follows to provide detailed compara- tive data. Laoma leimonias (Gray, 1850). Description of soft part*. Foot and tail slender, rounded pos- teriorly, not tapering noticeably. Sole undivided, without corruga- tions i fig. 25d). Pedal grooves high on foot, suprapedal very much smaller than pedal, uniting above tail. Pedal groove passing around tail with a distinct constriction on top of tail. No protruding caudal horn, but top of tail indented, with subpedal region extending pos- teriorly. No middorsal groove. Slime network very inconspicuous. Body roundly truncated anteriorly. Body color in preservative yellow white, ommatophores black. Surface of pallial roof and visceral hump with frequent black speck- les, less frequent white marks. Mantle collar (MC) elongated (fig. 25a), with bluntly rounded anterior edge. Basal margin with long intrusion of mantle gland onto pallial roof, corresponding with position of large left anterior mantle lobe (MA I at parietal-palatal angle. Anus (A) opening at inner edge of mantle collar (fig. 25b), slightly anterior to external urinary pore (KX). Urinary chamber (LK) a sharply defined groove that is im- mediately united with the V-shaped anal channel passing forward in the mantle collar. Pallial region (fig. 25a-b) very elongated, extending P/i-2 whorls apically from aperture. Basal Vi whorl with glandular exten- sion of mantle collar (MG). longest where major palatal lamella is formed and reabsorbed. Kidney (K) U-shaped, pericardia! and rectal branches equal in length. Rectal arm of kidney lapping completely under hindgut (HG) and extending onto parietal margin of pallial cavity (fig. 25c). Thus, rectal arm lies on both parietal and upper palatal margins of pallial cavity. Pericardia! arm of kidney appears more slender in Figure 25a, but solely because pericardium lies un- derneath kidney. Hindgut (HG) paralleling parietal-palatal margin well past apex of pallial region. Loop of intestine (I) with apical base KD ^-^\ Q """ "p^~" HV 2 mm 1mm x ^ ^r 2 GD FIG. 25. Anatomy of Laoma Icimonian iGrayi. Herekino Gorge, Kaitaia, Northland, North Island, New Zealand. L. Price! XI-1962. FMNH 135401: a, pallia! region; b, detail of pneumostomal openings; c, detail showing extent to which kidney