EFFECTS UPON FISHES OF CHANGES IN SALINITY OF WATER. 93 



studies. The following experiments, however, show a lii<ili range of varial)ilit\- in 

 till' individuals used — much higher than in JF". (lid/i/ininii^." Hence the results an- 

 somewhat less decisive. Owing to a scarcity of material, it was found necessar\- to 

 use Hshes of very ditlerent sizes, hut this fact certainly does not account Un- the 

 variations in the proportion of chlorine. 

 E.eperiinent 108. 



Marone from Tashmoo Pond (1.0U2) kept for some days (unfed) in water of 

 same density in laVjoratory before ))eing used for experiment. One small speci- 

 men used for each test, {a) Percentage of chlorine, O.llH; (/-) percentage of 

 chlorine, 0.135; (c) percentage of chlorine, 0.151; mean, 0.131. 

 This mean figure, it will be recalled, is identical with that for the /'. i]iaji/iJi,iii.< 

 from this same water. In view of the great range of variation iiere. howe\er. su<-li 

 pi-ecis(> agreement is doubtless due to chance. 

 Erpcrinient 109. 



MiifiDii' from Lagoon Pond (head of Vineyard Haven). A wattT saniplr 

 taken near shore gave a reading of I.ok;. but it is likely tiiat the water in 

 which the tisiies lived was considerably more saline. The fishes were kept in 

 the salt water of the laboratorj' (1.023) 1 and 2 days, respectively, before the 

 analysis. One fish was used in each experiment. («) Percentage of chlorine. 

 0.136; (J>) percentage of chlorine, 0.142; mean, 0.139. 

 The difference of this figure from that of the brackish-water specinuns is per- 

 haps not great enough to be significant. Strict candor compels me to record the 

 analysis of another fish from this source, which appeared to contain only o.lOl per 

 cent of chlorine. This figure falls so far below all others obtained from this species, 

 however, that I can not but regard it as due to an error. If this analysis were 

 included with the foregoing, the average woidd become 0.l2(i. 

 Experiment 110. 



Fishes from a fresh-water pond (landlocked). The water of tiiis pond, 

 though fresh to the taste, was found to contain 0.051 gram <'hloriiie per liter, 

 or about IS times the proportion in the fresh -water supply of New York City. 

 One fish was used in each test. («) Percentage of chlorine, o.ll2: (li) per- 

 centage of chlorine, 0.130; mean, 0.121. 

 Were there no other facts upon which to base conclusions, it might be objected 

 that the differences in salt content of these fishes from various sources might be due 

 to differences in their food and not to anj- osmotic relations between the animals and 

 the surrounding water. Such an objection is of course inapplicable to the fore- 

 going experiments with the various species of Fundulus. It is likewise out of ques- 

 tion as regards the next two experiments with Morvne. Here fishes coming originally 

 from brackish water were analyzed after a stay of some days (unted) in salt or fresh 

 water in the laboratory. 

 Experiment 111. 



Fishes from Tasiuuoo Pond were kept (> days in water of density 1.0:^3 

 before the analysis. In the second of the two specimens used the alimentary 

 canal was removed. («) Percentage of chlorine, 0.136; {}>) i>ercentage of chlo- 

 rine, 0.162; mean, 0.14:9. 



:)f the latterspeciesanuinbcr of .spt 



