352 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



rV row is suppressed, leaving 17 rows, which continue throughout 

 the rest of the body. 



A specimen such as just described would have the scale rows 

 recorded as — 



Neck, 19. Body, 19-21-19-17; 

 the sequence of addition and suppression of the scale rows as — 



Neck, —IV. Body, + V, —V, —IV; 

 and the complete data showing the gastrostege level at which the 

 added rows begin and the suppressed rows end may be presented as 

 follows : 



One of the variations in the scale formulae is where the body instead 

 of having 19-21-19-17 rows, has only 19-17, that is, 19 rows ante- 

 riorly and 17 posteriorly. In this type it is the fourth row in actual 

 counting which is lost. 



In order that the data be correctly coordinated the scale rows must 

 be given permanent numbers on that part of the body on which the 

 highest count occurs in the species. In this case it is 21 rows, and 

 the series are numbered from I to X in addition to the median. A 

 specimen in which the highest count is 19 rows must be assumed to 

 have the V rows suppressed constructively. It will have 9 rows on 

 each side of the median; for these to be recorded in terms of the max- 

 imum number for the species they must be counted as I, II, III, IV (V 

 suppressed constructively), VI, VII, VIII, IX, and the median row. 



Another variation is where the count is 19-17-15. In tliis type 

 the 19 rows become reduced to 17 by the suppression of the IV row, 

 and occasionally the 17 rows become 15 by the loss of the fifth row 

 in actual count, but this row when recorded in terms of the maximum 

 rows for the species must be counted as the VI row. 



Behind the posterior pair of geneials there are usuaU}^ one or two 

 pairs of small gular shields; these are followed by from one to three 

 shields in the median hue which increase in width in pyramidal 

 fashion. The shield that is regarded as the first gastrostege is the 

 first one that is nearly the standard width; it is usually distinguished 

 by being colored similarly to the rest of the ventrals and not white or 

 cream-colored like the throat. 



VARIATION IN NUMBER OF DORSAL SCALE ROWS. 



Combining the records obtained by Doctor Ruthven^ with the data 

 secured from the present series it appears that there are eight dis- 



1 Bull. 61, U. S. Nat. Mus., p. 148. 



