282 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



hind margin reaches more and more posteriorly 

 with respect to a vertical through the posterior 

 margin of the eye, as the size of the fish increases. 

 This is important to bear in mind, since species 

 are sometimes distinguished on the basis of the 

 position of the posterior margin of the maxillary 

 with refei'ence to the position of the eye. Again, 

 the relation of the maxillary length is usually ex- 

 pressed as the number of times it enters into the 

 head length. Since both the head and maxillary 

 decrease relatively with size, at least up to a cer- 

 tain point, the numerical value of their ratio does 

 not change much with the size, except in speci- 

 mens under 50 mm. Anybody desiring to use this 

 ratio may readily compute it by simple division 

 of the averages given in the tables. However, 

 this ratio does not always express specific differ- 

 ences, as for instance in the case of tvoolmani and 

 hrasiliensis. In these two species the measure- 

 ment of the maxillary as compared with the 

 standard length, shows a pronounced divergence; 

 but when the maxillary length is compared with 

 the head length the divergence disappears, 



SPECIMENS AT THE BORDER LINE 



Inspection of figures 1 to 4 afford sufficient 

 proof that the three common eastern species are 

 distinct. Since, however, there is often more or 

 less intergradation when any single specific char- 

 acter is considered, it is of some importance and of 

 considerable interest, to consider in greater detail 

 how specimens at the border line were referred to 

 their proper species in constructing the tables and 

 graphs as presented in this report. After all, in 

 identification it is individual fish that we are deal- 

 ing with, and in such closely related species it is 

 important that individual specimens are referred 

 to other proper species. 



First of all, it may be pointed out that the 

 number of actually overlapping specimens are 

 very few insofar as it relates to the counts of the 

 gill rakers and the anal rays. In the case of the gill 

 lakers (tables 3 and 4) there are no intergrading 

 individuals between dentattis and alhigutta or 

 lethostigma. In the case of the anal rays (table 

 5) there would be no intergrades between alhi- 

 gutta and lethontigma or dentatus if only two in- 

 dividuals each of the latter two species are elimi- 

 nated from the 3S1 specimens counted. 



However, the number of actually overlapping 

 specimens is not of primary interest. It is of 

 greater interest to know, in such closely related 

 species, just how all other specimens near the bor- 

 der line have been pi-operly referred. For in- 

 stance, two specimens having a total of If! gill 

 rakers have been referred to dentatus. Wliat is 

 the reason for placing them in that species and 

 not in alhigutta. since as far as the frequency dis- 

 tribution of that single character is concerned, it 

 would be just as logical to refer them to the latter 

 species (compare with table 4). Of these two 

 specimens one has D. 89, A. 70, scales 67, and the 

 other has D. 95. A. 71, scales 64. The color pat- 

 tern is also that typical of dentatn-?. It is evident, 

 therefore, that these other characters unmistak- 

 ably remove these specimens from alhigutta. They 

 apparently belong to dentatus and are extreme 

 specimens with respect to the gill-raker count. In 

 the same way, other specimens at the border line 

 with respect to any character may be referred with 

 confidence to the proper sjiecies by at least one 

 character falling outside the range of the most 

 closely related species and at the mode or even the 

 extreme outer end of its species. 



Infrequently, no character is entirely decisive, 

 but one character is sufficiently pronounced that 

 the specimen may be placed with assurance. The 

 following two fish from North Carolina are ex- 

 amples of such specimens. One has D. 83, A. 64: 

 gill rakers 13-1-4, scales 63 ; the other, D. 80, A. 61, 

 gill rakers 14 + 4, scales 59. They have ocellated 

 spots, but the color pattern is somewhat intermedi- 

 ate between detitatus and alhigutta and not typical 

 of either species. It will be noted that in the anal 

 ray count the former specimen is more like denta- 

 tus and the latter more like aJh/gi/tta. The scales 

 in the first specimen fall somewhat outside the 

 range of alhigutta, and considering also that tlie 

 number of anal rays is just outside the range of 

 that species, it would be more properly placed 

 with dt'iitatiis. In the second specimen, the num- 

 ber of scales falls at the beginning or at the end 

 of the frequency distributions of the two species, 

 resjiectively. Both have been placed with dentatus 

 laigely on the basis of the gill-raker count. Ref- 

 erence to table 4 and figure 4 will show that this 

 is the proper disposition of these two specimens. 



After boi-(ler-line individualssuch as the preced- 

 ing ones are placed, there remain a few speci- 



