l8 DISCOVERY REPORTS 



in the juvenile fish it is doubtful whether they are functionally efficient since Waterman found that in 

 his specimen of the Ceratioid Gigantactis longicirra, which was probably a juvenile, the structure of the 

 slender tubular nostrils and their innervation suggested to him that olfaction was unimportant. 



Table 2. Ceratias holbolli Kroyer. Growth constants (a) of the dimensions 



Part 



(a) Length of illicium 



(b) Length of exserted part of basal bone 



(c) Total length of basal bone 



(d) Length of caruncle, stalk included. Specimens 1-7 

 Length of caruncle, stalk included. Specimens 8-12 



(e ) Snout to base of tentacle 

 (/) Snout to dorsal fin 



(g) Snout to anal fin 



(h) Snout to base of pectoral fin 



(/') Length of maxillary 



(_/) Length of lower jaw 



(k) Greatest depth of body 



(/) Depth of caudal peduncle 



(m) Base of tentacle on head to base of tentacle (papilla, pore) on back 



(«) Dorsal fin to caudal fin 



(0) Anal fin to caudal fin 



(p) Base of tentacle (papilla, pore) on back to dorsal fin 



(q) Posterior caruncle to dorsal fin 



(r) Base of dorsal fin 



(s) Base of anal fin 



(t) Diameter of eye 



(«) Length of nostril 



Value of the 



growth constant 



(a) in y = bx a 



078 

 0-97 



1-20 



-0-34 

 0-98 



0-99 

 i-oo 



0-98 



0-90 

 0-91 

 1-05 

 0-92 

 0-96 

 i-oi 



0-98 



0-92 



I -00 



1-04 



1-04 



o-6 



o-8 



It remains to consider the exceptional dimensions b and d. There is obviously no lineal or other 

 relationship between the scattered points obtained on the log log graph of dimension b, the length of the 

 exserted part of the basal bone. In view of the adjustability of the lure, this scattering is precisely what 

 one would expect. Any appearance other than a random scattering would have seriously damaged 

 Bertelsen's arguments for invalidating the genus Mancalias. 



Both Parr and Bertelsen have stated that the relative size of the caruncles decreases strongly with 

 increase in length of the individuals. Actually the growth of these curious organs is somewhat more 

 complicated, and the comments of these authors require modification. The log log graph of caruncle 

 length (d) shows that growth of this organ in small individuals (nos. 1-7) is not negatively allometric as 

 Parr & Bertelsen implied. Rather there is marked positive allometry (a=i-2o) at this stage, and it is 

 noteworthy that the young caruncle is the only organ or dimension in the growth of Ceratias holbolli 

 which shows significant positive allometry. 



A slight error in the measurement of a comparatively small dimension like caruncle length becomes 

 magnified when plotted on a logarithmic scale of this kind. Remembering this it is seen that five of the 

 six remaining points can be represented by a second straight line. The last point (no. 13) refers to the 

 caruncles of Kroyer's type and is an approximation calculated from the somewhat vaguely delimited 

 caruncle in the figure of the type in Gaimard (1842-56, Poissons, pi. ix). Under these circumstances it 

 would appear reasonable to ignore it. The caruncles of all the specimens are now accounted for. 



At the point of intersection of the two lines, the growth constant changes abruptly to a=-o-34. 

 This change in sign of the growth constant means that at a period corresponding to an individual of 

 1 2 cm. and of caruncle size 075 cm. (Fig. 5 D) the growth of the caruncle changes abruptly from positive 



