Bijdragen tot de Dierkunde, 50 (1): 105-144 — 1980 Regression model evolution as exemplified by the genus Pseudoniphargus (Amphipoda) by Jan H. Stock Institute of Taxonomie Zoology, University of Amsterdam, The Netherlands Abstract A forma described in 1955 by S. Karaman, was raised to specific rank by G. Karaman in 1978. The has been considered genus Pseudoniphargus long mono- In the I have revised of the present paper, most specific. Its unique species, Ps. africanus, was supposed to materials attributed to the as well occur both sides of the the Atlantic existing on Mediterranean, on genus, Iberian Ma- side of the peninsula, on the Azores, and on as a good number of newly collected samples. This deira, in localities from the shore than ranging sea to more nine named resulted in a classification into species, 1000 m of altitude, and covering almost the entire natural and a number of unnamed left-overs because of salinity range (0-36‰). A taxonomie revision revealed that at least nine named insufficient material. The actual names for the species and several unnamed forms (of which insufficient taxa of Pseudoniphargus, formerly usually recorded material is available) hide under the name Ps. africanus, Ps. shown in table each with and as africanus, are I. a narrow ecological geographical range. The evolutionary scenario of the members of the is Like in the Gammarus identification of genus group, discussed of marine at some length: they are presumably females is much more difficult than that of males. origin, and got adapted to conditions of continental waters both sexes should be available for a during various marine regressions in the Eocene and Oligo- Preferably cène, but notably in the Miocene. profitable study. Much of the "variability" recorded in the has been based past on secondary 1. Contents sexual differences within one species. The evolution of — devel- 2. Introduction 105 Pseudoniphargus the 105 3. Acknowledgements opment of inhabitants of inland waters from 4. model evolution Regression (generalities) . 107 marine ancestors — is discussed at some length. 5. Evolutionary scenario and zoogeography of Pseudo- It is that this evolution is of the niphargus 108 likely regression 6. Cladistics 113 model type: stranding of marine populations 7. Place of Pseudoniphargus at level 114 family . during sea-level regressions. 8. Taxonomy 116 The has been undertaken to 8.1. Genus Pseudoniphargus Chevreux, 1901. present study com- Diagnosis 116 the ideas forward pare previously put (Stock, 8.2. Key to the species 116 model ideas 1977b) on regression evolution, 8.3. Description of the species 117 based obtained in the West 9. References 142 mainly on evidence Indies, with data derived from the Mediterranean marine area where theories on the stranding of 2. INTRODUCTION populations were first developed. Since the erection of the genus Pseudoniphargus Chevreux in almost student of it 3. ACKNOWLEDGEMENTS by 1901, every has stressed its "morphological variability" and its The author is indebted the and to following persons insti- striking distribution, from the sea shore to caves tutions for providing information or material: at considerable altitudes, from the South European Curators Lennart Cederholm and Hugo Andersson, Zoolo- giska Institutionen, Avdelningen för Systematik, Lund, Swe- and North African continents to oceanic islands den (abbreviated ZIL). like the Azores and Madeira. Dr. E. L. Bousfield, National Museums of Canada, Museum the remained of Natural Sciences, Ottawa, Canada. Up to 1955, genus monotypic. 106 J. H. STOCK - REGRESSION MODEL EVOLUTION Table I Previous records of Pseudoniphargus in chronological order. Unless otherwise stated, all records have been published under the name of Ps. africanus. Name in Author, year, page(s) Origin present paper (t = material re-examined; * = inferred name) Chevreux, 1901: 211-216, figs. 1-2 Bône (= 'Annaba) (Algeria) Ps. africanus Medjez-Sfâ (Algeria) Ps. macrotelsonis Schellenberg, 1937: 224 CantabrianMts. (Spain) Ps. elongatus Ps. unisexualis 1939: Bône Ps. Schellenberg, 297-304, fig. 1, 1 map (Algeria) africanus Medjez-Sfâ (Algeria) Ps. macrotelsonis Mustapha, Algiers (Algeria) t Ps. sp. Cantabrian Mts. (Spain) Ps. elongatus Ps. unisexualis Split (Yugoslavia) Ps. adriaticus Funchal Ps. (Madeira) t sp. Schellenberg, 1943: 1 Cost?, da Arrâbida (Portugal) Ps mateusorum Schellenberg, 1951: 327-328 Bonifacio (Corsica) Ps. adriaticus Balazuc & Angelier, 1951: 309-312 Banyuls (France) Ps. adriaticus S. Karaman, 1953: 139 Dubrovnik (Yugoslavia) Ps. adriaticus 1953: 672 Ps. Ruffe, Philippeville (= Skikda) (Algeria) sp. Balazuc, 1954: 184-185 Banyuls (France); Bonifacio (Corsica) Ps. adriaticus 44-51 Budva Ps. adriaticus S. Karaman, 1955: 236-241, figs. Dubrovnik, (in text) or Bar (in legends) (as formaadriatica) Dahl, 1958: 6-9, fig. 1 Faial, Pico (Azores) Ps. brevipedunculatus Ruffo, 1960: 172-174 Menorca (Balearic Is.) Ps. adriaticus Strinati & Coiffait, 1961: 228 Menorca (Balearic Is.) Ps. adriaticus Sket, 1969: 148 Rupine near Sukosan (Yugoslavia) Ps. adriaticus Margalef, 1970: 170-171, figs. 1, 2 J-L Basque country (Spain) Ps. unisexualis I Pi- sp. Mateus & Afonso, 1974: 10-11 Santa Maria (Azores) Ps. brevipedunculatus Ginet, 1977: 174-175 Cantabrian Mts. (Spain) Ps. elongatus Mateus & Mateus, 1978: 16-19 Costa da Arrâbida (Portugal) Ps. mateusorum Prov. Beira (Portugal) Ps. longispinum Prov. Estremadura (Portugal) Ps. Sp. Madeira Ps. sp. Azores Ps. brevipedunculatus G. Karaman, 1978: 250-256, figs. VI-IX Bône ( = 'Annaba) (Algeria) Ps. africanus Oued-ed-Demane (Tunisia) Ps. sp. Medjez-Sfâ (Algeria) Ps. macrotelsonis de Cueva San Adrian near Cegame (Spain) Ps. unisexuali s Do.: 241-250, figs. I-Y (as Ps. adriaticus) Budva (Yugoslavia) Ps. adriaticus Mola di Bari (Italy) Ps. adriaticus Isola di Montechristo (Italy) Ps. adriaticus Funchal (Madeira) Ps. sp. Menorca (Balearic Is.) Ps. adriaticus Mustapha, Algiers (Algeria) t Ps. sp. Prov. Santander (Spain) Ps. elongatus Split (Yugoslavia) Ps. adriaticus Dubrovnik (Yugoslavia) Ps. adriaticus Sukosan Rupinenear (Yugoslavia) Ps. adriaticus Costa da Arrâbida (Portugal) Ps. mateusorum Pesce Ps. Bari Ps. et al., 1979: 8 (as sp.) (Italy) sp. BIJDRAGEN TOT DE DIERKUNDE, 50 (l) - 1980 107 the Mme. N. Coineau, Laboratoire Arago, Banyuls-sur-Mer, models, original distribution area of a taxon France. is divided, the barrier being small at the beginning, Dr. J. Forest, Muséum National d'Histoire Naturelle, Labo- but effective at a later model is ratoire de Zoologie, Arthropodes, Paris, France (MNHN). 100% stage (this Prof. R. Ginet, Département de Biologie Animale et Zoologie, particularly useful in understanding the develop- Université Claude-Bernard (Lyon-1), Villeurbanne, France ment of populations isolated under influence of (UL). Dr. H.-E. Gruner, Zoologisches Museum der Humboldt Uni- plate tectonics). versität, Berlin, D.D.R. (ZMB). A good regression model taxon (genus, Prof. A. Mateus, Universidade do Porto, Instituto de Zoolo- family...) must satisfy the following five condi- gia "Dr. Augusto Nobre", Porto, Portugal (IZAN). tions: Curator G. E. Maul, Museu Municipal do Funchal, Madeira. Mr. Jos Notenboom, Biospeleo Werkgroep, Wageningen, The (1) In addition to the species in continental Netherlands. waters, marine relatives must be known. Dr. S. Pinkster, Zoologisch Museum, University of Amster- 2 should occur in areas once covered dam, The Netherlands (ZMA). ( ) They only travel of the University of Amster- above the surface. Furthermore, a grant by sea, but actually dam is (1979) gratefully acknowledged. should outside sub- (3) They not occur formerly remarks made model elements trans- The on regression in merged areas, since then dispersal from continental gression areas are based on unpublished observations made financed ruled during an expedition to the southern Bahamas, by areas to areas emerging cannot be out. the Netherlands Foundation for the Advancement of Tropical (4) They are lacking, or at least exceptional, in Research (WOTRO), The Hague, and the Beijerinck-Pop- areas that show ping Fonds, Amsterdam. transgression (subsidence). If is (5) there more than one inland taxon, these being developed through independent passive isola- 4. REGRESSION MODEL EVOLUTION tion marine the inland tend from ancestors, taxa (GENERALITIES) to have cladistic relationships indicative for the model and for the active Several authors (e.g. Bonnet, 1956; Hubault, passive (vicariance) not model 1938; Fryer, 1965) have emphasized the prob- (dispersal) (see Rosen, 1976, fig. 20). These five be adstructed follows: ability that populations of littoral or shallow-water points may as in inhabitants of the Ad The marine relatives should animals, more particular 1) by preference interstitia of stranded to the same the limnic macroporous substrates, be- belong genus as species, but do have be of the cause of various marine regressions in the Medi- they not per se to ancestors limnic since the marine forms have terranean region during the Tertiary epoch. These species, may stranded the further evolution after the isolation of populations got gradually adapted to undergone conditions of more continental waters and are now the limnic taxa. Ad 2 Several this distributed in land areas that were previously sub- ) examples prove to a surprising the distributionof the Thermosbaenacea merged. extent, e.g. Stock (1977b), studying groundwater Amphi- (Fryer, 1965; Stock, 1976), of Typhlocaris (Fryer, of the poda of the hadziid group in the Antilles, accepted 1965),