268 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



as the starting point of the correlation. It con- 

 sists of two somewhat irregular polygons which, 

 considered independently, might also be taken to 

 represent two species. Were the number of gill 

 rakers of the specimens represented by the right 

 polygon graphed separately, the result would be a 

 bimodal polygon similar to their figure 79, which 

 would represent two species, lethostigma and 

 denfafus; while the left polygon of figure 80 rep- 

 resents aJhIgutta. A comparison of Hiklebrand 

 and Cable's figures 79-81 with figures 1-3 of this 

 report will clarify the preceding discussion. The 

 intraspecific variability and distribution of the 

 three characters concerned, among the three com- 

 mon species, are such that when a mixture of speci- 

 mens of the three species is studied and the mixed 

 data graphed for each character separately, as 

 was done by Hiklebrand and C;ible. the resulting 

 polygons would be similar to their figures 79-81, 

 leading to the conclusion that not more than two 

 species are involved. But when the characters are 

 correlated it becomes clear that three distinct and 

 common species are represented. Moreover, after 

 correlating the characters and dividing the mass 

 of specimens into three rather well-defined species, 

 other characters appear which although not suffi- 

 ciently divergent to separate all the specimens will 

 yet distinguish the great bulk of specimens of the 

 three species, respectively. 



Norman ^ states : '■'•{alblgutta is] perhaps identi- 



cal with P. lethostigma . 



(p. 75 



and 



". . . it is possible that lethostigma, alhigutta and 

 squamilentus will eventually have to be regarded 

 as representing one variable species" (p. 76). 

 However, when adequate sanqiles of the tliree 

 species about which Norman was in doubt are 

 studied and the data correlated and tabulated, as 

 is done in tlie following pages, all questions as 

 to their distinctness disappear. While Norman 

 tentatively did treat these three species as dis- 

 tinct, he did not properly separate all his western 

 Atlantic specimens.- 



In order to prove that the separate species are 

 distinct, and to show how individual fish may be 



^A systoinntic nionop:raph of the Fl.itfishes (Heterosoinata) 

 vol. 1, Psi'ltuilidac, Botliiilao. Plueronectidae, by J. R. Norman, 

 British Muscmn, London. ]li:;4. 



' .See Gin.sl>Hrg, .Tour. Washington Acad. Sci., vol. 26. pp. 

 130-133. 1936. In that paper I discuss briefly some of the 

 differences between the present treatment of the species and 

 that in Norman's work. Where necessary the discussions are 

 here amplified under the accounts of some of the species. 



identified, it becomes necessary to investigate the 

 chief distinguishing characters by statistical 

 methods ; in other words, it is necessary to deter- 

 mine in detail the variability of these characters 

 of each species separately, showing precisely their 

 limits and their normal frequency distributions, 

 and to correlate them. That has been accom- 

 plished during the present study for the com- 

 mon species, as far as available material permits. 

 It now becomes a comparatively easy matter to 

 separate the species. There is seldom trouble in 

 placing individual specimens, certainly not more 

 so than in many other closely related species. 



A study such as that reported in the following 

 pages manifestly must precede any consistent 

 study of the life history of each species. Besides 

 studying their taxonomy, the known and scat- 

 tered data regarding the biology and the econom- 

 ics of the species have been digested and con- 

 densed, and original observations included. This 

 pa]ier treats of those species that inhabit the At- 

 lantic and Pacific coasts of North and South Am- 

 erica. The species are so closely interrelated that 

 it is necessary to treat them as a group in order 

 to understand them fully. 



In stating proportional measurements of cer- 

 tain parts throughout this paper, the figures given 

 refer to percentage of the standard length. State- 

 ments of the size of specimens refer to the total 

 length, including the caudal fin. Measurements 

 of the eyeball and orbit are those of the ujiper 

 eye. The stated number of scales refers to the 

 number of rows over the straight part of the lat- 

 eral line unless otherwise specified (p. 271). The 

 diagnoses include only those characters which are 

 of importance in distinguishing the species. 

 Counts and relative proj)ortions are mostly given 

 in general statements in the diagnoses. More 

 detailed data are in the tables which form part of 

 and should be used in connection with the diag- 

 noses. 



In the following accounts of the species, the 

 given numbers of specimens e.xamiiied are (hose 

 in the United States National Museiun Catalog, 

 unless otherwise indicated. 



All illustrations accompanying this paper, ex- 

 ecuted with such obvious skill, were prepared by 

 Louella E. Cable. Figures of specimens represent 

 reworked photographs, which were made in the 

 Smithsonian photographic laboratory. 



