94 BULLETIN OF THE KUREAU OK FISHERIES. 



This figure is somewhat greater than that for the fishes taken from a somewliat 

 (liluted sea water (experiment lOit). It will be noted that the second fish gave a very 

 much higher percentage of chlorine, in spite of tiip rcmoviil of the ;ilimont:irv 

 canal. The contents of the latter, therefore, played no part in deterininiiii;- tiie 

 amount of chlorine present. 

 K.'prniHvnt Ihi. 



Fisiies from Tasluuoo Pond, kept 5 days in the fresh water of the lahoia- 



tory pri(jr to analysis. In the first case. 1 fish was used, in the second -1. 



In />, the alimentary canals were removed, {n) Percentage of chlorine, ti.li^o; 



(J>) percentage of chlorine, 0.109; mean, 0.114. 



This figure will be seen to be somewhat lower than that for the fisiies from tiie 



fresh-water pond (experiment 110), but in view of such large variations it would lie 



absui'd to draw any conclusions from this fact. 



If, on the one hand, the mean figure for the Monme from the fresh-water pond 

 is averaged with that for those kept in the laboratory fresh-water for 5 days: and if. 

 on the other hand, the figure for the Vineyard Haven fishes is averaged with that for 

 those kept in the laboratory salt water, the figures thus obtained, together with thai 

 for the brackish-water fishes, may be arranged in the following series: 



( (J ) Fresti-water specimens 0.117 



(6) Brackish-water siieeimens I). i:!4 



(<•) Salt-water siieciiiieiis 0. 144 [nr 0. KiT] 



The figure for the "fresh water" fishes is thus 1!> per cent [or 15 per centj l(>ss 

 than that for the "salt water" ones. It was found in experiment 97 (also with Mororu) 

 that O.OI)*') gram chlorine per 100 grams body weight was given out in the course of a 

 single da}'. Assuming the percentage of chlorine originally present to have l)een (i. 1 W 

 it would appear that 25 per cent of the chlorine of the l)ody was lost in a single day. 

 Here, then, the figures obtained by the two methods are not in full accord. 



It may be regarded as abundantly proved by the preceding experiments that 

 some species of fishes, at least, undergo considerable changes in the salt content of 

 their bodies in consequence of changes in the salinity of the surrounding water. 

 These changes are fairly rapid, resulting, at times, in difierences of 25 per cent or 

 more in a single day. Moreover, they do not necessarily result in any harm to the 

 animal. It is plain, however, that these changes in the bodily salt content are in no 

 way proportional to the changes in the salinity of the medium. Indeed, very great 

 alterations in the latter may sometimes be made without any apparent effect upon the 

 former. (Experiments 103, i, and lo7.) The exact extent of the correlation between 

 the two might be determined Ijy a sutticient number of analyses, but it is quite 

 unlikely that it could be expressed by any single mathematical fornuda. This is 

 because fishes do not conduct themselves as simple dialyzers. 



Whether or not all of the tissues are e(pially affected b}- these changes in the 

 bodily salt content I can not say. The bod\' as a whole loses or gains in its propor- 

 tion of salts (chlorides), and these leave or enter it by some path other than the 

 alimentary canal. The occurrence of the weight changes described in the preceding 

 section seems to show the existence of membranes which are permeable in some 

 measure to water, at least. In the absence of any other plausible hypothesis it ma}' 

 t)e assuuied that the salts enter or leave the body through these same membranes. 



