RESEARCHES ON RACES OF HERRINGS. 81 
mediate stages are met with between a trace of an internal cross-piece 
and a complete arch. It is not improbable that in the living animal these 
arches are closed by a cartilaginous cross-piece. Cursory examination of 
the records indicates that in a majority of skeletons the sum of the number 
of vertebree with complete hemal arch and the number with well- 
developed (i.e. potentially complete ?) arches is 33; or it might be said 
that the commonest number of vertebrae with potentialities for complete 
heemal arches is 33; the greatest number of such vertebrae appears to 
be 35. 
EXPLANATION OF CHARACTERS 13, 14, AND 17 IN 
SAMPLE IV. 
In Number IV Sample it was thought that more information could be 
obtained by analysing the dorsal and anal fin (Characters 13 and 14) ina 
slightly different way from that adopted in Sample III. 
Thus in Sample IV all the anterior fin-rays of the dorsal fin which were 
distinctly shorter than the longest fin-ray were counted separately from 
the following and recorded in the general form of a+-b, where a@ 1s the 
Fic. 4.—Dorsal Fin. In Sample III the Fic. 5.—Anal Fin. In Sample III the 
above would be recorded as 2+17; in above would be recorded as 2+15; 
Sample IV as 3+16. in Sample IV as 3+14. 
number of the smaller anterior rays and b the number of rays posterior 
to these. In this way it is possible to reconstruct a fin to represent the 
commonest form which by cursory examination of the records is seen to 
be one having 3-16 rays. A drawing of the type of ray is shown in Fig. 
In the case of the anal fin (Character 14) the fin was analysed in the 
following manner : all the anterior fin-rays which were not subdivided at 
the tip or splayed out in any way were counted and recorded separately as 
“qa” from those in which the rays were splayed out—recorded as ‘‘ b.”’ 
The commonest form of fin is seen from the records to be one recorded as 
3+14. This type of fin is shown in Fig. 5. 
Character 17 in Sample IV is also recorded in a manner slightly different 
NEW SERIES.—VOL. XI. NO. 1. MARCH, 1916. F 
