84 PRELIMINARY EXPERIMENTS ON 



individuals from time to time had strange fits of torpidity, and at a 

 stage when the larvoe in H were already settling on the bottom, those in 

 F showed a marked preference for the surface, which they refused to 

 leave. They were obviously not quite healthy ; and, although I adopted 

 a different interpretation later on (see p. 86) in connection with similar 

 phenomena in jar H, I attributed the unhealthiness of the larv£e at the 

 time to the unsuitability of the conditions in the plunger-jar, since after 

 the larvffi were transferred to the stagnant water in H they assumed 

 habits similar to those of their companions almost immediately (next 

 day). 



It is possible, though unlikely, that the ill health of the later larvee in 

 F was due to a deterioration in the reservoir water, which was changed 

 a few weeks after the commencement of the experiment. In August 

 the harbour water was doubtless of an exceptionally pure character, in 

 consequence of the long dry summer, but the rains of September must 

 have subsequently afiected its salinity and quality in the neighbourhood 

 of the Laboratory, as they undoubtedly drove away the Eed Mullet 

 which had previously been lurking about the mouth of the Cattewater 

 in exceptional abundance. 



Thus in spite of the low mortality in this experiment during the first 

 few weeks, which points to the suitability of the harbour water at that 

 time, I should hesitate to conclude that the harbour water would be 

 usually suitable for the development of this fish in other years or at 

 other seasons of the year, when the salinity is lowered by the greater 

 quantity of fresh water coming down the Tamar, Tavy, and Plym. 



Experiment G requires little to be said of it in consequence of the 

 initial defects in the jar already alluded to. On the nineteenth day 

 (September 4th), prior to my Channel cruise, I preserved the four 

 survivors in this experiment in order to have some material for study- 

 ing the anatomical details of the progress made in development. The 

 larvte varied in length at this stage between 6"8 mm. and 8*0 mm. The 

 maximum depth through the hinder part of the eye was 2"0 mm. in the 

 largest specimen. The notochord was turned up posteriorly in the 

 largest, but scarcely bent at all in the smaller specimens. Two hypural 

 bony plates were developed in the largest, but were quite insignificant 

 in the smallest larva. Delicate caudal fin-rays in fan-like tufts were 

 conspicuous in all, and formed a protruding ventral lobe to the fin in 

 the smaller specimens. 



A noteworthy, though not conspicuous feature of tlie larva up to this 

 stage, and even for another week beyond it, is the permanently gaping 

 mouth. The larvae do not close their jaws until they are a full month 

 old, when the teeth begin to appear. 



Experiment H was started on August 22nd, in order to see the effect 



