BULLETIN OF THE FLORIDA STATE MUSEUM BIOLOGICAL SCIENCES Volume 2 Number 3 A TAXONOMIC REAPPRAISAL OF THE TURTLE PSEUDEMYS ALABAMENSIS BAUR Archie Carr and John W. Crenshaw, Jr. UNIVERSITY OF FLORIDA Gainesville 1957 The numbers of THE BULLETIN OF THE FLORIDA STATE MUSEUM, BIOLOGICAL SCIENCES, will be published at irregular intervals. Volumes will contain about 300 pages and will not necessarily be completed in any one calendar year. William J . Riemer, Editor Roland F. Hussey, Associate Editor All communications concerning purchase or exchange of the publication should be addressed to the Curator of Biological Sciences, Florida State Museum, Seagle Building, Gainesville, Florida. Manuscripts should be sent to the Editor of the BULLETIN, Flint Hall, University of Florida, Gainesville, F16rida. Published 15 August, 1957 P'rice for this issue 8.30 A TAXONOMIC REAPPRAISAL OF THE TURTLE PSEUDEMYS ALABAMENSIS BAUR ARCHIE CARR AND JOHN W. CRENSHAW, JR.1 During·a recent study of certain sections of the emydine turtle genus Pieudemys, Crenshaw made a search for new character5 that might be brought to bear on the question of relationships among the complexes in the southeastern United States. One aspect of the study led to a re-evaluation of the placement of Pseudemys alabamensis Baur in the synonymy of Pseudemys /101-idana mobilensis by Carr (1952). 'The presen t paper proposes that, in spite of the fact that all its characters occur individually with greater or lesser frequency in one stock or an- other of the P. Boridana group, alabamensis may be a real genetic en- tity occupying a continuous small range in the Gulf drainage of the southeastern coastal plain. If so, its retention in synonymy can only lead to the ignoring of a zoogeographic situation which, though by no means well under5tood, appears to be of extraordinary interest. Crenshaw was successful, as will be shown below, in finding at least one character that seems as good as, or better than, any hitherto used in defining the Pseudemys rubriventris group. A reasse&sment of alabaniensis in the light of the new criterion brings out or re-empha- sizes four salient facts: (1) The diagnosis of alabamensis, formerly weak, is ·strengthened. (2) The characters of alabamensis suggest P. nelsoni of peninsular Florida as the nearest relative. (3) P. alabamensis appears tot form with nelsoni, and with rubrivent,·is of the Atlantic coastal plain north of North Carolina, a species complex with a curi- ously discontinuous distribution-a catena of stocks so closely knit that the gaps geographically separating the three forms seem incon- gruous. (4) P. alabamensts shares with ne/soni and iubriventris, and perhaps accentuates, a peculiar relationship with the #oridana group, a section of the genus which, when judged by typical examples only, seems so.divergent as to be of almost subgeneric level. 1 Archie. Carr is 1?rofessor of Biological Sciences, University of Florida, Gainesville. He is the author of several books ; among them are Handbook of Turtles - and Guide to the Reptiles, Amphibicrns and Fresh -water Fishes of Florida (with C . Goin) John Crenshaw, Jr..is Assistant Professor of Biology, Antioch College, Yellow Springs, Ohio. His dottoril .dissertation is concerned with the genus Pseudemys in the southeastern United States. He gralefully acknowledges the support of the De- partment of Zoology, University of Missouri, where he was employed during the preparation of this manuscript. Manuscript Submilted 19-August 1956.-ED. - 26 BULLETIN FLORIDA STATE MUSEUM Vol. 2 TAXONOMIC HISTORY OF PSEUDEMYS ALABAMENSIS Baur (1893: 224) described alabamensis as a distinct species of the genus Pse*demys and noted that it was closely allied to rubriventris. He pointed out that the form had been treated by Agassiz (1857:433) under the name mobitensis. Baur designated syntypes (without cata- logue numbers) in the collection of Mr. G. Kohn of New Orleans; these were deposited subsequently in the c611ection of the United States National Museum. Siebenrock (1909:469) suppressed alabamensis to a synonym of rubriventris, but Stejneger and Barbour (1917: 120) ignored his action and treated it as distinct. Carr (1938:106, 107), noting the almost concurrent ranges of alabamensts and the mobilensis-suwanniensis complex of the species Boridana, and the agreement in coloration and conformation shown by "alabamensis" from various 10calities and the phase of Boridana charactefistic of those same localities, suggested that alabamensts might represent no more than an occasional variant of Boridana. About the same time, Stejneger (1938:173), in clearing up certain errors of identification in the literature, referred to ala- bamensis as a subspecies of P. Tubriventris; but subsequently he re- turned to his earlier opinion and treated alabamensis as a distinct species (Stejneger and Barbour, 1943:203). Carr (1952:280, 303) treated alabamensis as a synonym of P. Boridana mobilensis for the reasons given above, and his treatment has been followed! by sub- sequent workers. INTERGROUP RELATIONSHIPS Three more or less distinct groups of turtles of the genus Pseudemys occur in the territory under discussion: (1) the well-differentiated Scripta group with no known morphologic overlap with either of the others; (2) the rubriventris group, composed of ,·ubriventris, ala- bamensts, and nelsoni; and (3) the #oridana group. Four forms meet in the territory occupied by alabamensis-the coastal strip from the panhandle of Florida to New Orleans . The turtles of the Boridana group, in themselves zoogeographically puzzling, are dealt with by Carr in previous papers and most recently by Crenshaw in a detailed analysis now in manuscript form (Crenshaw, MS). The difficulties of interpreting relationships among the three major Floridian stocks, and among the many variant populations composing them, are bad enough when they are considered separately. Now, however, it ap- pears necessary to evaluate them in the light. of what appears to be 1957 , CARR AND CRENSHAW : PSEUDEMYS 21 rather widespread genic interchange, or at least genic community, with the rubriventris group. Crenshaw (MS.) has traced out and evaluated several examples of character duplication as it bears upon the taxonomy of the Boridana group; other aspects of their relationships, bearing upon the status of P. alabamensis, must be examined here. 15 NELSONI and P. F. PENINSULARIS P. F. SUWANNIENSIS Figure 1.-Generalized distribution of three turtles of the genus Pseudemys in peninsular Flotida. P. f. peninsularis is not actually known to extend south of the broken line across southern Collier and northern Dade Counties. PSEUDEMYS NELSONI AND THE FLORIDANA GROUp.-The range of P. nelsoni and the combined-ranges of P. #oridana peninsularts and P. f. suwanni- ensis are nearly coextensive (fig. 1), with ' even the ecologic position of nelsoni being apparently transgressed by the others. Because of this, one is at first inclined to VieW the markedly divergent appearance of adult ne/soni, of the "typical" phenotype, as the mark of a "good 28 BULLETIN FLORIDA STATE MUSEUM ,Vol. 2 species" in the conventional sense. It iS therefore doubly surprising to find that in occasional specimens, and in some populations, the dis- tinction between the two groups weakens or breaks down completely. While a number of characters have been used in differentiating between nelsoni and the Boridana stocks, only two seem to be consis- tently diagnostic. These involve the conformation of the anterior face of the upper beak and a feature of the head pattern brought out for the first time by Crenshaw (MS). In peninsularis and suwanniensis the tomium of the upper jaw is entire, anteriorly, whereas in nelsoni there is a prominent symphyseal notch accentuated by a toothlike cusp on each side (fig. 2). With re- spect to head markings, in all three forms a supratemporal stripe passes from the dorsolateral .area of the neck, anteriorly across the temporal region, thence above and immediately adjacent to the eye, and out along the snout toward the external nares. This stripe is often obscure in large nelsoni. Another pair of prominent stripes, the paramedians, are situated between the supratemporals, one on each side of the median line. In helsoni the paramedian stripes usually pass fr6m the neck anteriorly across the occipital region to terminate abruptly posterior to the level of the eyes. lIn suwanniensis the para- medians typically continue anteriorly from the neck, running between the eyes to a point well out on the sncut. Some specimens of penin- sularts show the same pattern, but more often the paramedians pass anteriorly from the neck to a point posterior to the ey.e where each bends laterally to join, or almost join, the adjacent supratemporal. Unfortunately, both of the above characters may break down in very young ne/soni which may exhibit both a smooth upper tomium and a head-stripe pattern typical of suwanniensis. The reduction of head striping characteristic of ne/soni seems to be correlated with increase in size, and it is possible that notch and cusp development may also be ontogenetic, although this has not been demonstrated. , Two additional characters now appear to be consistently useful in defining nelsoni. In suwanniensis and peninsularis the supratempo- ral head stripes either do not join anteriorly or the junction js faint and the stripes pass through or immediately adjacent to the nostrils (fig. 2). In nelsoni the supratemporals pass forward from the eye and join conspicuously well above and posterior to the external nares. A sagittal stripe pas5ing anteriorly from between the eyes usuall¥ joins the supratemporal stripes at their confluence to form a prefrontal arrow-shaped figure (fig. 2). Of 79 specimens of ne/sont examined, only 14, all medium-sized or large specimens, had cne or more breaks in the pretrontal arrow. The pattern was suggested in every instance. ]957 CARR AND CRENSHAW: PSEUDEMYS 29 To determine applicability of the character, a total of 45 P. rubri- ventris from Virginia, West Virginia, and New Jersey were examined. All of these individuals had a prefrontal arrow except five in which - it was replaced by a prefron tal transverse stripe. Another had a P. F. SUWANNIENSIS P. F. P.ENINSULARIS P. NELSONI -- 11** c:-Conform- Head Striping Second Lateral Lamina ~ ~ ~ Figure 2.-Cmnparison of three Florida turtles of the genus Pseudemys with respect to upper jaw conformation, head striping, and color· pattern of the second lateral lamina (left side). slight break in the arrow pattern, and three with generally obseured head markings exhibited very fairit arrows. Apparently, the pre- frontal arrow is characteristic of the rubriveniris group, and absent in the Boridana complex. The prefrontal arrow is, found also in the Pseudemys seripta group, but its frequency has not been studied. The other character found to contribute toward intergrbup diag- nosis involves the color pattern Of the second lateral (6r "costal") 80 BULLETIN FLORIDA STATE MUSEUM Vol. 2 lamina of the carapace. It is primatily of aid in the separating of netsoni and suwanniensis, but also distinguishes between some speci- - mens of netsoni and peninsularis. In the last two forms the basic pattern of the second lateral consists of a prominent, centrally located, light, transverse (vertical) stripe or band (fig. 2). This light band passes from the ventral (lateral) edge of the scale toward the apex of the dorsal (medial) edge. Just short of the apex it usually bends sharply posteriad and passes to the posterodorsal edge of the lamina. There may be another branch passing posteriad from the anterodorsal edge of the lateral lamina to join, or almost join, the posterior arm of the main transverse stfipe, thus forming a ¥-shaped figure. Rare variations include transverse bands which pass dorsad (mesiad) through the apek, or stop abruptly just short thereof without bending: In penin- sularis there are usually two to six narrow, transyerse stripes parallel- ing the central band ; in adult netsoni such accessory stripes are Often lacking and, even if present, are faint and few in number. Juvenile nelsoni usually exhibit a number of accessory stripes which. apparently fade· with age. Iii suwanmensts the basic pattern of the second lateral lamina is different from that of both nelsoni and peninsularis. As in the last two, a central, tran5verse light stripe may pass dorsad from the ventral (lateral) edge ok the lateral lamina, bend posteriad short of the apex, and thence run to the posterodorsal edge (fig. 2). More important, a longitudinal stripe passes from the middle of the median stripe to the posterior edge of the lamina. The C-shaped figure thus formed by the upper half of the median stripe and the two horizontal stripes-one from the middle of the median stripe; one from the upper end-is virtually always present in suwanniensis. The lower half of the median stripe may be lacking, and other stripes may or may not pass from the outer bend. of the C-shaped figure to any edge of the lateral lamina. Concentrie light circles are often present within the C-shaped figure. Thus, there are four characters which may be employed in dif- ferentiating among these three forms. Two of these, the pre- frontal arrow and the notch and cusps of the upper jaw, are chafac- teristic of nelsoni but rarely present in peninsularis and suwanniensis. A third character, the color pattern of the second lateral lamina, is of 'primary importance in recognizing suwanniensts, but is also of some value in distinguishing between nelsoni and peninsularis. Finally, although the head stripe patterns of all three forms differ on the whole, the pattern usually found in suwanniensts occurs too frequently in the other two to be of much help in the determination of that 1957 CARR AND CRENSHAW : PSEUDEMYS 81 form. However, the head stripe patterns described for peninsula, is and netsoni occur but rarely in other forms and acc6rdingly are of some diagnostic value. Another feature of particular use in differentiating between su- wanniensis and peninsularis, is the more or less symmetrical and ex- tensive pattern of dark markings· always evident on the plastron of suwanniensis. In peninsularis dark plastral markings are usually lacking, and even when present they are rarely symmetrical and never extensive. While nelsont may exhibit an immaculate plastron like peninsularts, or approach suwanniensis in regard to distribution of plastral markings, it typically lies somewhere between the two. Its plastral markings are usually markedly asymmetrical and not so ex- tensive as in suwanniensis. RELATIONSHIPS OF PSEUDEMYS ALABAMENSIS.-At the time Carr pro- posed relegating alabamensis to synonymy, his acquaintance with eusp- jawed Pseudemys from eastern Gulf drainages had involved six live individuals from New Orleans shown him by Percy Viosca; two live topotypes from Mobile in a lot of a dozen "Mobilians"-a local name for P. f. mobilensis plus P. alabamensis-plus a half dozen topotypes, including the type (fig. 3), in the collections of various museums; and two hatchlings "from brackish water" at Crystal River On the Gulf Coast of Citrus County, Florida. The point that stands out in Carr's notes is that, except that they all had notched jaws, the "alabamensis" at New Orleans looked strikingly like the local versi6n of the P. #ori- dana group (P. vioscana Brimley); those at Mobile resembled mobilen- sis; and the Crystal River animals seemed suspiciously similar to su- wanniensis. When these bbservations were made, Carr knew nothing of the character proposed by Crenshaw-the oblique prefrontal stripes joining anteriorly to form, with the sagittal stripe, the prefrontal arrow. Moreover, a re-examination of the little Crystal River speci- mens (UF 1736) shows that they may well be the young of typical P. nelsoni. Besides this material Carr had seen an6ther specimen, the evaluation of which, even in the light of the resurrection of alabamen- sis, remains a puzzle . A most surprising occurrence of the alabamensis phenotype is seen in an individual (ANSP 4521) from "Wolf River, Shelby County, Tennessee": no other example of a notch-jawed Pseu- demys can be found from anywhere near this region. This unique specimen-more recently examined by Crenshaw and found to show the alabamensis head pattern-contributed toward the rationale on which Carr based his earlier decision to suppress alabamensis as some sort of spurious phenotype. -*i 1.4'il.~r :..4 L :i ~9#0...'u:EF:*· 2*,.,.-0, .4 * , Gl I.* * 1 f 1.0 Figure 3 .-Dorsal and ventral views of the holotype of PS€U demys alabamensis ( USNM 20966) photographed prior to the 0 . 14- ,1 4 1 . . 4' :,L . . . .4 - . i* - , 44 + , *m„* '. I ..4., .44 i 71 1/W-, - 12741& - ,. .4, :t,*.drtz:. :.4 A 5 .*I. 153 54 '0.:156¥ -AL .P. 4 I ./ removal of the head. Note the dark plastral vermiculations which are also characteristic of adult malcs of both rubriventris and nelsoni . 34 BULLETIN FLORIDA STATE MUSEUM Vol. 2 Crenshaw has recently re-examined the holotype (USNM 20966), the only paratype (USNM 20967), 6 topotypes (USNM 55501 and 55736; MCZ 1660-1663) and 2 Floridian specimens of atabamensis-one from near Apalachicola (USNM 101090) and one from Wakulla Springs, Wakulla County (Crenshaw collection 1019-1). The skin of the heads of the holotype and paratype is now lacking, but in all the other specimens 'the prefrontal arrow is evident, while in all specimens Of P. f. Botidana and P. f. mobilensis, which are, respectively, the eco- geographical vicariants of peninsularis· and suwanniensis, the pre- frontal arrow is lacking. Moreover, the pattern of the second lateral in all specimens of alabamensts was basically like that of nelsoni and different from that of mobilensis (which is identical with that of , suwanniensts). The subspecies P. f. #ondana, which also occurs in the Mobile Bay area, is like peninsularis with respect to all the characters described above-except that the pattern of the supratemporal and para- median head stripes is like that of suwanniensis. The basic color pattern of the second lateral lamina of #oridana is thus similar to that of alabamensis. Differences in the color patterns of the lateral laminae of the Mobile Bay examples of alabamensts and #oridana are evident but are not so pronounced in preserved material as in fresh material. Table 1 summarizes some of the diagnostic characters for the six forms discussed above. DISCUSSION It appears unwarrantable to withhold taxonomic recognition from alabamensis. If there actually is such a form, the criteria at hand draw it readily into the rubriventris group and name as its nearest ally Pseudemys nelsoni of the Floridian peninsula. The only character known to distinguish consistently between the two is the section of the head pattern described above. In nelsoni the paramedian stripes end back of the eyes (fig. 2), but in alabamensis they continue anteriorly between the eyes and onto the snout as in suwanniensis. Reasoning from. the facts thus far presented, a fairly familiar zoo- geographic design emerges. If intergradation between the northern Gulf coast alabamensis and the peninsular Florida nelsoni were es- tablished, and if the gaps separating the ranges of these forms from that of rubriventris-the coastal plain of the northeastern United States*-were partially filled in, the characters common to the three would be taken as evidence that they are a f6rmenkreis with geo- graphic discontinuities due to incomplete collecting or to recent ex- tinction at the hand of man. 1957 CARR AND CRENSHAW: PSEUDEMYS 35 TABLE 1. SUMMARY OF SOME CHARACTERS OF SOUTHEASTERN REPRESENTATIVES OF Pseudemys alabamensis, nelsoni, AND #oridana. + al ab a en sis OSIO c cp o ·J s!icIns Dd · J 8!SU O I O w .3 + STS 0 ! eitns ·J Inhabits "panhandle" of + western Florida Itihabit5 peninsular F.lorida + + + Notch and cusps on upper jaw + + Prefrontal arrow + C-shaped figure on second + + lateral laminae Dark plistral marks +or-+or- + + Parallel, rare + 6ccas. +,+not reduced Supratemporal Parallel, and para- paramedians occas. + median head reduced stripes are: · Joined ~ posterior occas. + to eyes In 1942 Carr called attention to concurrence in geographic varia- tion between rubriventris populations and those of the Boridana group. Recent work by Crenshaw has turned up additional evidence of phenotypic duplication and suggests that the parallelism should not be attributed to independent genes similarly selected, but may be based on actual kinship. The reason such an explanation is hard to accept i5 the fact that 36 BULLETIN FLORIDA STATE MUSEUM Vol. 2 the two groups occupy the same territory and the same series of habi- tats, and the thought of their being interfertile or even being as closely related as the consonance would imply goes against accepted logic. We realize of course that barriers can be unseen as well as seen; but even so, it is not easy to visualize a relationship that allows two species to show so much evidence of genic interchange and still retain modal integrity in a common habitat. The distribution of characters within the #oridana group alone is complex enough, with its multidirectional clines, its arfay of slightly differentiated stocks isolated to varying de- grees by barriers both ecologic and geographic, and intergrading in some places but remaining sympatrically discrete or vicariating without intergradation elsewhere. But when phenotypic overlapping with the rubriventris group is added to this picture, the confusion of variables acquires a third dimension, and it becomes even more difficult to piece together the zoogeographic history of the animals. A good place to illustrate this problem is in the Apalachicola drainage of western Florida. There are two forms of the #oridana group in this area. One of these is the.local version of the mobilensis-to- suwanniensis cline; and the other is close to what is known as P. f. Boridana._ The relationship between the two seems to conform to the pattern for the complex, of which a chief feature is the intergrading of "subspecies" in some sections of their adjacent ranges and their remain- ing disjunct in others. In the Apalachicola Valley a complicating element is superimposed on this pattern when one sees in the Boridana population an emergence of what seems a strOng tendency toward the characters of Pseudemys nelsoni of the rubriventris group. We have four young specimens (UF 440-443) from "between Scotts Ferry and Blountstown," immediately to the north of the known range of alabamensis (taken by a collector in the employ of Ross Allen and presented to us by Wilfred Neill) that are indistinguishable, by any characters known to us, from netsoni of the peninsula. The agree- ment is so close that it was hard to believe that a mistake had not been made in assigning the data. Conversation with Neill and Allen corroborate the locality however. Moreover, the small number of other specimens in hand from the area and the large series of live specimens casually ekamined at the Ross Allen Reptile Institute of Silver Springs, Florida show decided tendencies toward nelsoni and connect the four young specimens with the average. of the local #ofidana stock in 'all respects except for the notched jaws and prefrontal arrows of the former. In spite of its similarity to nelsoni, the #oridana population is consistently smooth jawed. In this charac- ter and in the j6ining of the clear-cut prefrontal stripes, all four of 1957 CARR AND CRENSHAW : PSEUDEMYS 87 the little turtles, as implied above, agree more closely with ne/soni than with the nelsoni-like floridana stock in their own area. The hatchlings may then be (1) real P. ne/soni from nelsoni parents, which would extend the range of that species northwestward some 150 miles, (2) variants from the modal condition of the local alabamensts or floridana stock resulting from chance mutation or recombination, or (3) representatives of a derne: of the local alabamensts or #oridana stock that our inadequate sampling from the area has not otherwise turned UP. In any event, consonance of the specimens with the local trend is interesting, and they call attention to the fact that Apalachicolan #oridana are like nelsoni in the following ways: (1) Lack of a reticulate pattern with dominant C-shaped marks on the lateral laminae. (2) Emphasis 6f a vertical light band across each lateral, with obsolescence of other light markings on the laterals. (3) Deepening of the dark ground color of the carapace to nearly black. (4) Variability in plastral markings, with adult specimens most often unmarked below. (5) Occurrence of intensely pigmented smudges on only a few inframarginal seams instead of rings or ocellated spots on most or all of the seams. (6) Abrupt termination of the paramedian head stripe behind the eye with reduction of head pattern generally. (7) Erythrism of the lower shell. The most extraordinary implication of this situation-with or without reinforcement by the four questionable hatchlings-is that Boridana of the Apalachicola drainage, and indeed all #oridana populations, are probably genetically much closer to the rubriventris group than any previous study has indicated. There is, in fact, no known character of the rubriventris group that fails to turn up some- where, in the Boridana complex. As an example of the sort of detailed "give-and-take" that occurs, the color pattern of the marginal laminae may be offered. The upper marginal pattern of netsoni comprises three extremely different sets of markings, as follows: (1) a predominantly ocellate pattern involving a C-shaped figure with the open side directed pos- teriorly and enclosing a light spot, an inverted L-shaped figure an- terior to this,. and a broad stripe along the lower edge of the lamina; (2) a pattern composed of three centrally located longitudinal stripes parallel to the same broad stripe involved in the preceding and thus 38 BULLETIN FLORIDA STATE MUSEUM Vol. 2 made up entirely of longitudinally oriented lines; and (3) a single conspicuous median vertical light band running from the marginal stripe toward or to the upper angle of the lamina and separating it into two fields that may be unmarked or with scattered irregular mark- ings. Selected typical examples of these three styles of patterns show great divergence indeed-much stronger, for instance, than the diagnostic divergence of many named races. And yet all three occur in sets 6f nelsoni siblings (e.g., UF 1168, six hatchlings from Lake County, Florida) in both typical form and in every possible intermediate combi- nation; some specimens have all marginals marked with one or an- other of the patterns, while in others there is bilateral or antero- posterior replacement of one pattern by another. This range of variation involves such marked differences and such complicated intergradations that it is astonishing to find, on t6p of all the other parallelisms, that in the area between northwestern Marion County and the base of the panhandle, Boridana populations often show precisely the same spread in pattern type. When this coincidence in seemingly irrelevant and nonadaptive features is added to the ex- amples of intergroup concurrence in variation already known it be- come5 increasingly difficult to see how the hidden isolating fact6rs in these two sympatric stocks could possibly be operating to maintain the measure of disjunction that exists. Although the present study has not included the extensive re- appraising of material outside the southeast, certain known features of extra-Floridian sections of the Boridana group clearly bear on the problem of character-sharing by this and the rubi-iventris group. Even the notch-and-cusp-the most trenchant distinctive feature of the rubi iventris group-is diagnostic also of the westernmost representative of the species #oridana, viz., P. f. texana of western Texas and ad- jacent parts of Mexico. The fact that this race actually represents the geographic and morphologic extreme among Boridana turtles can be seen in the clinal nature of annectent hoyi, which has, not notched and cusped, but emarginate upper jaws. It has not been pointed out else- where that another conspicuous feature of texana-a tendency for one of the lateral head stripes to break up and reorient its tympanal seg- ment to form a conspicuous spot or vertical band through the ear- can be seen in both notched- and smooth,jawed turtles from Eoastal Louisiana and Alabama (e.g., MCZ 1659). Yet another character, of unkn6wn significance, but involved in the problem, is the peculiar "marbled shell"-carapace and plastron evenly mottled by the disrupted and completely redistributed pattern 1957 CARR AND CRENSHAW : PSEUDEMYS 89 of pigmentation-found not only in a small percentage of rubi iventris and nelsoni but also in the holotype of alabamensis. While no marbled specimen 6f Bo,-idana is now at hand, Carr has seen the tendency among tuftles being hauled away from Lakes Iamonia and Jackson, Leon County, Fiorida, when they have gone dry in years past; and the late Leonhard Stejneger told him of having seen it in material sent by a collector from thd same. localities. Carr has recently re- examined two of these specimens (USNM 95765-766). Both seem to him intermediate between nelsoni and Foridana. One (95765) even has a notched upper jaw, although no cusps are present. The pre- frontal arrow is lacking. THE QUESTION OF NOMENCLATURE If Pseudemys nelsoni ranges northwestward into the Apalachicola area (fig. 4) as a discrete genetic entity, then one might attribute the gs ..6 - 7, -- o P. alabamensts • £. nelsoni Figure 4.-Distribution of Pseudemys alnbamensis and P. ne/soni. ne/soni-like traits in the #midana populations there to occasional hybridization, or, as Crenshaw (MS.) proposes for similar populations elsewhere, to introgressive evolution. Moreover, if the notch-jawed Mobilians (alabamensis) ranging eastward to the Apalachicola area and slightly beyond likewise represent a separate form and not just an irregularity in the Boridana matrix, and if alabamensts is, as it 40 BULLETIN FLORIDA STATE MUSEUM Vol. 2 seems, of rubriventris "blood," then one might conclude that ala- bamensts and nelsont are a single varying species with an apparent distributional problem due merely to' inadequate sampling. With the same line of reasoning the closely similar, but geographically separate, rubriventris falls into the design. A conventional nomenclatural pie ture of this situation would be shown by trinomial listing of the three, with the specific name derived from LeConte's P. rubriventris, which. has priority. But the situation is not quite so simple when the detail of our evi- dence is considered. With respect to the relationships of the north- eastern ):itbriventris to the other two forms, we know only that they are separated by an apparent broad hiatus across the states of Georgia, South Carolina, and North Carolina. We have not studied the mor- phological Variation in rubriven Cris quantitatively and cannot state with any assurance. that it is, or is not, genetically distinct. In view of the frequency with which interspecies reproductive barriers are transgressed in nature within the genus Pseudemys, we do not feel that probability of biological reproductive isolation is a criterion of usefulness in the present situation. The relationship of nelsoni- to alabamensis is not nearly so clear as it might be. From the "panhandle" of western Florida, lying between the principal ranges of alabamensis and nelsoni, we· have a sample of six specimens, all with reliable collection data (fig. 4). Two of these specimens, both taken near the coast but from different localities, are typical of adult alabamensis. The other four, taken from a single locality some distance inland and north of the coast, are equally typical of juvenile nelsoni. Our small and manifestly inadequate sample suggests that ala, bamensts and nelsoni overlap geographically in the Apalachicola area, behaving as perfectly good species, perhaps by virtue of microgec- graphic or ecological isolation. In view of the single kn6wn charac- ter difference between the forms, it seems possible that we have sampled an intergrading population and that chance alone has determined that our four specimens from one locality be unlike the two available froni other 16calities. When a trinomial is used, if it means anything at all, it means that a situation has been pretty well worked out, is understood, and is nothing to worry about any m6re. It should give grounds for a sense of security that in the present case would be wholly uncalled for. Look- ing only at rubriventris turtles, the move might be logical in spite of the geographic gap. Appraised against the background of the extra- rubri©entris relations of the notch-jawed stocks, the situation seems 1957 CARR AND CRENSHAW: PSEUDEMYS 41 to call for a more noncommittal disposition. After some bickering be- tween ourselves, therefore, ·we propose to eschew the trinomial-of- complacency and reinstate the notch-jawed Mobilian under a binomial- of-reserved-judgment, Pseudemys alabamensis Baur. CONCLUSIONS In the light of new criteria we propase the rehabilitation of Pseu- demys alabamensis as a member of the turtle fauna of c6astal areas of Alabama, Louisiana, and northwestern Florida, and perhaps of eastern Texas. Through one contiguous or sympatric relative or another, alabamensis appears to intergrade with stocks of both the rubi iven.tris andthe Boridana groups. It may be that the much deplored morpho- logical overlap of these turtles, which have confused naturalists for more than a hundred years, can best be exemplified by, and may eventually be largely explained through, 5tudy of the "alabamensis" phenomenon, which, whether a "good species,- a sporadic variant, or merely a clumping of endpoints in several geo- and ecoclines, is bound to repay any attention given it. ACKNOWLEDGMENTS For assistance in a number of ways, and for permission to examine specimens in their private collections or in their care, we are grateful to the following: L. N. Bell, University of Miami, Florida; Charles Bogert, American Museum of Natural Histbry (AMNH); Doris Coch- ran, United States National Museum (USNM); Roger Conant, Phila- delphia Zoological Gardens; J. C. Dickinson, University of Florida Collections (UF); Arthur Loveridge, Museum of Comparative Zoology (MCZj; Neil Richmond, Carnegie Museum; Jay Savage, Academy of Natural Sciences, Philadelphia (ANSP); and Albert Schwartz, Charles- ton Museum. To save space we have not listed the location and catalogue numbers of specimens examined, except in the case of alabamensis and of, other material which seemed critical or possibly conttoversial. It may be of help to future workers to note that series of netsoni are available in the Carnegie Museum, the Charleston Museum, and the University of Florida Collections. Most of the rubri- ventris material examined is from the Carnegie Museum, the Museum of Comparative Zoology, the Philadelphia Academy, and the private collection of Roger Conant. A considerable number of people have helped in collecting, have provided specimens, suggestions, criticism, or varied abettance. To ' all of these we are deeply grateful. We are especially obligated to - 42 BULLETIN FLORIDA STATE MUSEUM vol. 2 Claude Adams, Walter Auffenberg, Fred Berry, David Caldwell, J. C. Dickinson, Coleman Goin, Arnold Grobman, Keith Hansen, Robert Hellinan, Richard Highton, William Jennings, David Karraker, John Kilby, Ellis Lanquist, Ann Lyons, Larry Ogren, Beth Schultz, and C. H. Wharton, all, now or formerly, of the University of Florida; to Lewis Marchand, Gainesville, Florida; to Albert Schwartz, Charleston Mu- seum; to L. H. Babbitt, Petersham, Massachusetts; to Ernest Williams, Museum of Comparative Zoology; to Roger Conant, Philadelphia Zoological Gardens; and to William Duellman, University of Michigan. LITERATURE CiTED Ac.Ass:Z, Louis 18.57.. Contributions to the natural history of the United States. Vol. 1. Boston: Little, Brown and Company; li + 452 pp· BAUR, 6. 1893. Notes on. the classification and taxonomy of ihe Testudinata. Proc. Amer. Phil. Soc., vol. 31, pp, 210-225. CARR, ARCHIE F. 1938. A new subspecies of Pscudemys Bo,idana, with notes on the floridana complex. Copeia, no. 3, pp 105-109. 1942 . The status of Pscudemys Boridana texana, with notes on parallelism in Pseudemys. Proc. New England Zool. Club, vol. 1, pp. 69-76. 1952. Handbook of Turtles. Ithaca, N. Y.: Comstock Publishing Associates, xv + 542 pp., 37 fig., 82 pls., 23 maps. CRENSHAW, JOHN W.,JR. MS . The ecological geography of the Pseudemys floridana complex in :the south - eastern United States. Unpublished dissertation, Univ. Florida. SIEBENROCK, F. 1909.. Synopsis der rezenten Schildkr6ten. Zool. Jahrb., Suppl., vol, 10, pp 427-618. SiEJ NEGER, LEONHARD 1938 . Restitution of ihe name Plychemys hoyi Agassiz for a weStern river tortoise . Proc. Biol. Soc. Washington, vol. 51, pp. 173-176. STIC.~NEGER, LEONHARD and THOMAS BARBOUR 1917., A check list of North American amphibians and reptiles. First ·edition. Cambridge, Massachusetts: Harvard University Press, iv + 125 pp. 1943. A check list of North American amphibians and reptiles. Fifth ediliOn. Cambridge, Massachusetts: Harvard University Press, xix + 260 pp. Contributions to the BULLETIN OF THE FLORIDA STATE MUSEUM may be in any field of biology. Manuscripts dealing with natural history or systematic problems involving the southeastern U nited States or the Caribbean area are solicited especially. Manuscripts should be of medium length-12 to 200 printed pages. Examination for suitability is made by an Editorial Board. The BULLETIN is distributed worldwikle through institutional subscriptions and exchanges only. It is considered the responaibility of Lhe author to distribute his paper t6 all interested individuals. To aid in this, fifty copies are furnished the author without cost. PREPARATION OF MANUSCRIPT Manuscripts should be typewritten with double spacing, with ample margins, and on only one side of the paper. The author shou Id keep a copy; the copy 5ubmitted must be the original. Tables, legends of figures, and all footnotes should be assem- bled on sheets separate from the text. Several legends or footnotes may be placed on a single sheet. Illustrations, including maps and photegraphs, should be referred to as "figures" wherever possible. All illustrations are reduced to a makimum of 4-1/4 by 7-1/8 inches. The scale, wherever· it is necessary, should be incorporated. into the figure. All references to' literature should conform with the bibliographic style used in recent numbers 6f the BULLETIN. Spell out in full the titles of non-English serials. - Footnote material should be kept to a minimum. However, provide copy for a footnote detailing the title, affiliations, and address of the author (see recent nunl- bers of the BULLETIN). The scope of the paper should be stated in an introductory paragraph. A summary should be provided whenever appropriate. It should be complete in itself, and understandable without reference to the body of the article. Manuscripts and all editorial matters should be addressed to: Editor of the BULLETIN Flint Hail University of Florida Gainesville, Florida