EFi'ECTS UPON FISHES (IB^ CHANC4ES IN SALINITY OF WATER. 91 



(c) Had spent 1 day in fresh water. Percentage of chlorine 0.12tj. 



(d) Fresh water 2 days. Percentage of chlorine 0.129. 



It will be noted, first, that the lot from salt water showed a percentage of chlorine 

 not very different from that of J^. heterocUhis (preceding experiment). Second, the 

 percentage was practically the same in the fishes from water of density 1.001 as in 

 those from pure sea water; while third, fishes which had spent one day in fresh water 

 showed a loss of 30 per cent of their chlorine. Fourth, there was pi-actically no 

 difference in this regard between those wliich had been in fresh water 1 day and 

 tl)ose which had been there for 2 days. 



The second of the results just enumerated is x'ather surprising, in view of the 

 fact that F. heteroclitus was found to 3'ield up a considerable fraction of its chlorine 

 in water of approximately the same densit}' (1.001d=). though retaining it in water 

 of densit}^ 1.002. Whether this difference of results is due to an}^ real physiological 

 difference between these two species may be doubted. Where the water is so near 

 to what may be called the critical degree of salinit}', a slight difference in one direction 

 or tlic othei- niigiit suffice to produce a very considerable difference of results. It 

 was previously found (experiments 31, 31 h)is) that water of density 1.001 was slowly 

 fatal to F. laajalln, while water of density 1.002 effected little or no harm. 



Despite minor discrepancies, the last two experiments agree in showing a great 

 difference between the effects upon the chlorine content of the body of pure fresh 

 water and water having a cei-tain small percentage of salt. This difference is 

 extremely significant in view of the difference, already discussed, in their effects 

 upon the life of the fishes. 



In another series of experiments (lol- 107) F. <J)iii>hiiniix was used. These fishes 

 had been taken in Tashmoo Pond, in water of density 1.0u2, and kept in the labora- 

 tory unfed for a number of days prior to the experiment in water of the same 

 density. Analysis of this water showed that it contained 7 pei- cent as much chlorine 

 as the local sea water. The entire fishes were used for analysis, the alimentary canals, 

 etc., not being removed. The specimens were first rinsed thoroughly, here as alwavs. 

 Erj.rfwrnt 10\. 



Fi\ e diffcri'nt analyses were made with fishes taken directly fi-om the 

 brackish water (l.(»(»2). In each case 1 fishes were treated together. These 

 analyses gave the following percentages of chlorine: a, 0.175; A. 0.13(5; c\ 0.128; 

 d, 0.139; e, 0.132. The mean of these figures is 0.U2. 

 .It will be seen that with the exception of a these figures do not present a wide 

 variation. I regard the first result as probably due to an error, both on account of 

 the exceptionally large amount of chlorine indicated and because the test was not a 

 ver}' satisfactory one. Strict fairness compels its msertion here, however. The 

 mean of the other four figui'es is 0.134, and this figure, being in my mind the more 

 proliable one, will be used in subsequent conjparisons. The reader is at any time 

 free to substitute the other figure. 

 E.rperiiii<'nt 105. 



Here the fishes were kept in fresh water for a \arying perifnl prior to the 

 analysis. 



(a) Iday (4 fishes) 0.112 



(6) 3 days (4 fishes) 0. lOS 



(c) 11 days (8 fishes) 0.085 



