80 BULLETIN OK THK KTTREAU OF FISHEKIES. 



weighing, in y t-iscs out of :;. hulcccl, the lislics in tlu'sc -1 cases all remained in 

 perfect health for another i'4 iionrs. after whieh they were returned to tlieii 

 normal medium. Thereafter no deaths oceurri'd (with one pos,sil)le exeeption) for 4 

 days more, after which the lishes were no longer observed. The tishes used iiad 

 been "seasoned" for a long period in the laboratory, and a large majority would 

 survive indefinite transfer to sea water (see p. t!7). 

 Experhfients 69-7 1. 



Six scup, wei^lied separately, were found to have gained to a scarcely 



appreciable extent if weiuhed imniediately after death. If allowed to remain 



long in the water, liowe\ cr, the gain was much more pronounced. 



In interpreting the foregoing figures it must be borne in mind that these 



experiments were made during the summer months when metabolism in general was 



more active, and when the loss of weight through waste, in these unfed animals, was 



conseijuently more rapid. Owing to this fact, and probably, also, to the discharge 



of mattei's from the reproductive organs, such accurate determinations of gain or 



lo.ss in weight as were previously made were here found to be impossible. It will be 



seen by reference to the above notes that in no case has a gain been recorded where 



it would not have been expected according to h^'pothesis. On the other hand, loss 



or stationary weight is recorded in some cases where a gain might reasonably have 



been expected. The discharge, at any time, of eggs or of ovarian iiuid from one or 



more of the fishes in a given lot might well have produced some of these aberrant 



results. This, of course, was not likely to occur during- the winter and early spring. 



Even such allowances do not seem sufficient, however, in the case of Fuixhihis 



majalif, as the following experiments show: 



Exp,rhin„ix 72-76. 



Of five lots transferred to fr(\sh water two showed a considerable gain, but 

 ill both of these cases the Hshes had begun to die, and dead fishes rapidh' 

 iml)il)(! water, as has already beciii pointed out. In a third case a slight 

 (scarcely significant) gain is to be noted. In a fourth lot there was an initial 

 loss of about 1 per cent during the first 1:^ hours, followed bv a gain of 

 more than 2 per cent during the next 12 hours. But it is to be noted that 

 in this case the control fishes in salt water likewise showed a loss during the 

 first half day, followed likewise by a slight gain. In the fifth fresh-water lot 

 a loss was recorded at the end of the first day, as well as at the end of the first 

 half day, the total loss amounting to about 3 per cent, while that of the control 

 lot amounted to onlj' a trifle over 1 per cent. 



Of the five lots put into 1.001 water, three were found to have lost during 

 the first half day, and likewise throughout the entire day, the total loss 

 amounting in each lot to more than that shown by the control fishes in salt 

 water. The other two lots gained during the first half day, but lost during 

 the second. 



Of the control fishes, three lots lost, one remained stationary, while one 



first lost and then gained slightly. 



Had I begun my experiments in weighing during the summer and used Fmuhi- 



lus majalk at the outset, it is likely that I should have despaii-ed of discovering any 



significant changes in weight, and, indeed, it may appear to the reader that such rig- 



