EFFECTS UPON FISHES OF CHANGES IN SALINITY OF WATER. 81 



iuo;s as have just been recorded maj' well have been due to accident or error. 1 will 

 l)roinptly add that I can not myself satisfactorih- explain these differences in my 

 results. 



It must be uri;ed at the outset, however, that these aberrant cases in no way 

 invalickite the results of m^' earlier experiments. From the earlier series we are 

 forced to conclude that certain changes of weight may occur as the result of osmotic 

 action. Some of the later experiments show that under different conditions such 

 changes ma}' not occur. The only differences in the conditions, so far as I am aware, 

 relate (1) to temperature (20'^ to 22^ C. at Woods Hole, 3' to 15^- C. at New York) 

 and (2) to the pre\ious history of the Hshes used. Those employed at the New York 

 Aquarium had been living in water of density 1.007 to 1.012, those at Woods 

 Hole (i. e., the salt-water ones) in sea water of nearly or quite full strength. I can 

 not believe, however, that this was a siguiticant factor in the case. 



It is well known that osmotic pressure varies with temperature. In the present 

 ease, however, it seems to be a difference in the permeability of the membranes that 

 demands explanation. During the summer moiith-s metabolism in general is of 

 Lourse more active, and various chemical and physical changes are in progress which 

 are dormant during the colder season. It is not unlikelj-, therefore, that the osmotic 

 permeability of a ffsh, always dependent upon the physiological condition of the ani- 

 mal, should present considerable differences during the summer and winter months. 

 This l)eing granted, it remains to be asked just what changes in permeability would 

 account for some of these aberrant results. In certain of the foregoing experiments 

 with the three species of Fundulus, where a decided gain was to have been expected 

 in view of earlier results, a decrease was found to have occurred instead. Assuming 

 that this decrease is attributable to normal waste as in the case of fishes left in their 

 customary medium (or to some other irrelevant circumstance, see p. 80), how are we 

 to account for a stationary weight in a stronglj' hypotonic fluid? As has already 

 been stated (p. 70), a stationary weight in a medium which is not approximately 

 isotonic may mean either (1) that the membranes are impermeable to water; or (2) 

 tliat water and salts maj' pass through with equal freedom, in which case the condi- 

 tions for osmosis are wanting. The latter alternative does not necessarily imply a 

 freel}' porous state of the memt)ranes, such as obtains in the case of a filter. It is 

 quite conceivable that this indiscriminate permeability of the body surface should be 

 subject to the controlling influence of the organism, just as we know that the differ- 

 ential permeability, which renders osmosis possible, may be so controlled. I have 

 thus emphasized this second possibility, inasmuch as I regard the first one as being 

 thrown out of consideration in certain cases. Although no unquestionable change in 

 weight resulted in either F. majalis or F. diaphmius when transferred from salt to 

 fresh water, I shall later show that a decided change occurred in both cases in the 

 salt content of the bod}-. It will likewise be shown to be highly improbable that this 

 salt passed from the bod}' by way of the alimentary canal or the reproductive organs, 

 leaving as the only alternative a diffusion through one or more of the limiting mem- 

 branes of the body. 



In some other cases, on the contrary, the facts are best explained by assuming a 

 condition of complete impermeability, as in the case of F. //uijali-i when transferred 

 from salt to slightly saline water. Here no certain increase in weight occurs, and 



B. B. F. 1905—6 



