A MONOGRAPH OF THE SEA-SNAKES (HYDROPHIINA:). 171 
rest of the series barely twice as broad (see fig. 57). In Platurus, Emydocephalus, and 
Aipysurus they attain their maximum development, being more than three times 
as broad in the whole body length and very similar to the same shields in Colubrine 
terrestrial snakes. 
Numerically the value of these shields is of importance in distinguishing certain 
genera, but in closely allied species like those of the genus Distiva, the range of varia- 
tion in individuals is so considerable, and the figures of the specific ranges overlap so 
much, that the assistance to be derived from the number of these shields is decidedly 
limited. They are fewest in the genera Emydocephalus, Atpysurus, and Hydrelaps, 
being less than 200 ; most numerous in Distiva fasciata where they may exceed 500. 
The specific range of variation depends largely upon the numerical strength of the 
individuals available. If we exclude species but poorly represented numerically, the 
smallest range of variation is that met with in Distiva jerdoni (219 to 248), and D. 
viperina (235 to 267); on the other hand the largest ranges of variation are seen in 
Distira fasciata (376 to 531), D. torquata (310 to 438), and D. cyanocincta (280 to 397). 
The ventrals of most species are entire or mostly entire in the whole body length, 
the few shields that are divided being seen about the umbilical scar and before the 
anus. InHydrus platurus,andin some specimens of Distira major, many of the shields 
ate divided, but subject to a good deal of variation in number and position in indi- 
viduals. In D.cantoris and D. gracilis all the shields in the posterior half or so of the 
body are very constantly divided (see fig. 13). In the very broad shields in some 
Emydocephalus and Platurus a median obtuse keel is seen posteriorly, but this is an 
inconstant feature found in only certain individuals, and, I believe, irrespective of age 
and sex. 
The remarks made with reference to keels, tubercles, etc., under costals apply 
equally to these shields. 
The ventrals in many of the species, specially in the genus Distiva, are often very 
difficult to count accurately. The difficulty may arise from the detail of these shields 
being obscured by damage, desquamation, a local sodden condition, or the puckering in 
places occasioned by the way the specimen has been folded in the bottle. Often too 
small scales are interpolated on one or other side, which would alter the count on the 
two sides. Some observers count these,some do not. Again some appear not to count 
the shields which may be broken up, especially those just before the anus, and others 
again do not count the early ill-developed ones in the neck. The result is that the 
counts of various authors for the same specimen differ considerably. To take a single 
instance, the type-specimen of Distiva cyanocincta has 308 ventrals according to Russell, 
290 according to Boulenger, more than 320 according to Gtinther, and I count them 
310. Itis not very unusual for me to make these shields a little different in three or 
four counts in the same specimen, which may appear extraordinary for one to confess 
who strives at accuracy : still it is the fact. 
CosTaLs.—(The ‘‘ scales’’ of other authors). The importance of these shields in 
classification is only second to that of the ventrals. 
