EFFECTS UPON FISHES OF CHANOES IN SALINITY OF WATER. 75 



(,i) Fishes ill (dilute) salt water (l.ooT to 1.009). Loss of 1 per cent in 5 

 days. Ill experiment 41, </, where normal tishes were used, there was a loss 

 of 1 per eent in 2 days, but the tishes had been much more recently fed than 

 ill the present experiment (here unfed Ki days), which may account for the 

 dirierence. 



{h) In fresh water (running). Gain of 2.s per cent in '1 tlays. 4 per cent in 

 4 days, after wliich tlie increitse ceased. Here the iiaiii for the tirst 2 days 

 is only slig'htly in excess of the gain in 1 day in the case of the normal tish 

 (experiment 41, h). 



(r) In l.uul water (running). Gain of 1.4 per cent in 2 days. (Unfortu- 

 nately the gain in 1 day was not determined). In (experiment 41. r. the tishes 

 gained 1 per cent during the tirst day, then lost. 



(i!) In l.nl4 water. Loss of 1.2 per cent in 2 days. This loss is not more 

 than is aceoimteil for liy waste. 



(() In 1.02.5 water. Loss of 4.3 per cent in 1 day; tj.l per cent in 2 days. 



In experiment 41, e, these losses were 4.1 and 8.3 per cent, respectively, but this 



is liardly a fair comparison, since in the latter case the water was for some 



reason harmful to the fishes. 



It will thus lie seen, by comparing the expei'imeiits ujion sound tishes with those 



in which the tishes were scrai)ed, that the removal of scales throughout a large area 



did not result in increasing the influx or efflux of water resulting from changes in 



the density of the medium. Garrev's contention (p. <i2) thus appears to be disproved. 



Indeed, later experiments seem to indicate that it is through the membranes of the 



gills and not through the general body integument that these osmotic changes mainly 



occur. 



The following experiment shows strikingly the difference between the effect of 

 pure fre.sh water aiul watei- which is slightly saline. It also illustrates the alternate 

 gain and loss of weight which a tish may ur.dergo in consequence of repeated changes 

 of density. 



Expei'lnuitf 1^6. 



New York, April. IKO.'i. Five specimens of the IS-spined sculpiii {^fyo.ro- 

 eej>/uiJi/f< octodec'i msjiriiosiix) taken from salt water of density 1.009. 



{i() Two specimens put directly into fresh water. After 24 hours lioth were 

 dead. Average increase in weight, 4.3 per cent. 



{h) Three put into 1.001 tank (density ranging from a half point above to a 

 half point below that figure). Average increa.se of 2.4 percent during Hrst 

 day; 8.8 per cent during 3 days. 



At this tiiu(> two of the fishes (Nos. 1 and 3) were Mck, the other (No. 2) 

 appearing quite well. It is to be noted that the gain in weight was (-onsider- 

 ably less in the case of this latter fish. The two former fishes w-ere accord- 

 ingly put back into the salt-water tank (now of density 1.011), No. 2 being- 

 left in the 1.001 tank. No. 2 gained an additional 2 per cent during the fol- 

 lowing day and died after 2 days more (6 days after beginning of experiment). 

 Nos. 1 and 3 underwent loss of 6.5 per cent during next 4 days (in salt water). 

 These last wei'e then put for a second time into l.Ool tank, when thei-e 

 resulted a gain of 4.8 per cent in 1 day. 



