RESEARCHES ON RACES OF HERRINGS. 121 
925 * In Character 17 the “a” portion is recorded hence first in order to 
avoid any bias towards making up the number to the apparent total 
of 33 or 34 potential arches. 
943 * 35th vertebra with two pairs of neural and hemal spines, nearly twice 
as long as normal, and apparently equals two fused. Counted as one. 
962 * Arches about middle of skeleton a little abnormal. 
975 * 12th vertebra nearly twice as long as normal, and apparently equal to 
two fused, but only counted as one. Also 30th vertebra with 13 
heemal arches and spines, but otherwise of normal size. 32nd also 
with abnormal heemal arch processes. 
980 * 39th vertebra with one hemal arch attached to arch of 38th vertebra. 
1010 * 37th vertebra has one extra neural and one extra heemal spine, but is 
otherwise normal. 
1064 * Vertebra behind Ist complete arch carries an incomplete but well- 
developed arch ; it is counted with those having complete arch. 
1067 * Left process of 44th hemal arch is joined up with junction of 43rd 
heemal arch. 
1075 * Ist hemal arch broken, but still attached to centrum. 
Note—The lateral processes of the vertebree preceding those with complete 
heemal arch are sometimes long without internal processes, and at other times 
shorter with the internal process developing. Nevertheless, the arbitrary 
character chosen has been adhered to as closely as possible. There are, 
however, doubtful cases. 
Very frequently the 4th, 5th, and 6th vertebree from posterior end have the 
base of the hzemal arch passing diagonally across the ventral anterior hali of 
the centrum, and in this respect are unlike the other vertebre. 
