276 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



line in the chart. In only one of the squares does 

 the broken line cross. That is, of tlie total number 

 studied only two specimens of alblgittta and one 

 of lethostigma have the same correlation of the 

 number of gill rakers and anal rays. Such speci- 

 mens must be assi<j:ned to their proper species by 

 means of other characters. 



In figure 4 is plotted the total number of gill 

 rakers. Practically the same result is obtained by 

 plotting the number on the lower limb only, except 

 that in that case the lines are more densely 

 grouped. 



FREQUENCY POLYGONS 



Three variable cliaracter.s wliich are of impor- 

 tance in distinguishing the three common species 

 from the east coast of the United States are rep- 

 resented graphically in figures 1 to 3. The poly- 

 gons representing the number of dorsal rays are 

 markedly irregular; those representing the gill 

 rakers are fairly regular; those representing the 



anal rays are intermediate with respect to regu- 

 larity of arrangement. The irregularities are ap- 

 paiently caused by imperfect sampling and may 

 result from one or all of the following main fac- 

 tors. (1) The number of specimens studied may 

 not be sufficient to form a representative sample 

 in its resjiective species. (2) The method of sam- 

 pling may be inadequate. (3) The samples do 

 not represent altogether homogeneous popula- 

 tions. It will be shown hereafter (p. 320) that 

 the populations of dentatus from Chesapeake Bay 

 and from North Carolina differ appreciably in 

 these three characters. To some extent this is 

 also true of diiferent populations of lethostigma 

 (p. 332), and probably also of alhigvtta, although 

 in the latter two species population differences are 

 appai'ently not so marked. The irregularities in 

 dentatus partly disappear when the data are 

 tabulated separately for Chesapeake Bay and 

 Nortli Carolina. 



Si- 55 56 57 5a 59 60 61 6Z 6i 6f 65 66 67 68 69 70 71 ^^. 73 7'f^ 75 76 77 



NUMBER OF AXAL RAYS 



Figure 4. — Corrplation between total number of Rill rakers on first gill arch and number of anal rays, of three common 

 east-coast species of PnmWehthija. (See p. 274). Each mark represents one specimen; four marks with a cross 

 line represent five specimens. 



