76 



SCIENCE 



[N. 8. Vol. XXXVI. No. 916 



ever, be made between men and women of the 

 same graduating classes. The average num- 

 ber of children of the married men in the same 

 classes as the women is 1.7 as compared with 

 1.34 for the women. The total number of 

 children of Wesleyan alumni, men and women, 

 excluding duplicates, is 4,731. Of these 1,076 

 have been reported dead. 



Frank W. Nicolson 



AN INDEX OF FISH ENVIRONMENTS 

 In studying the distribution and success of 

 fishes within a given area, one often notes the 

 absence of certain species from localities 

 which appear quite suitable fish habitats upon 

 inspection, or even upon determination of the 

 oxygen content of the water. Evidently, the 

 causes of this absence is not a life and death 

 matter such as is often supposed to be true 

 in animal distribution. But since fishes are 

 able to move readily from place to place, they 

 may leave or fail to enter a locality where the 

 conditions are entirely compatible with life 

 for a longer or shorter period. Tet the 

 change in conditions may be great enough to 

 cause the fishes either to turn back or to 

 leave the locality because of stimulation and 

 increased activity. Wishing to test this reac- 

 tion possibility, we devised a means of study- 

 ing the behavior of fishes when they encounter 

 difl'erences in gases or solids in solution. 

 The apparatus consisted of a device giving a 

 constant flow of water with desired amounts 

 of dissolved gases, and at any temperature 

 within ordinary experimental needs. Two 

 tanks, 120 cm. long by 20.5 cm. wide by 14 cm. 

 deep were arranged under identical and sym- 

 metrical surrounding conditions. Water was 

 introduced into both ends of the tanks at the 

 same rate and was allowed to flow out at the 

 center. The same kind of water was intro- 

 duced into the two ends of the control tank. 

 In the experimental tank the water intro- 

 duced at one end was like that of the control, 

 while the gas content of that introduced at 

 the other end had been experimentally modi- 

 fied. This established a gradient between the 

 two kinds of water. Fishes put into the 



tanks tend to go back and forth and thus 

 encounter the experimental gradient. When 

 the change of conditions thus encountered 

 was such as to affect the fishes, they reacted 

 either by turning back or by passing through 

 the gradient into the treated water. But in 

 this case they quickly returned to the un- 

 treated water, thus spending a shorter time 

 in the treated water. 



Eight species of fish, widely separated taxo- 

 nomically, were studied in detail. All the 

 fishes were slightly negative or indefinite in 

 their reaction to differences in oxygen con- 

 tent. We found no good evidence that they 

 react to nitrogen. Their reaction to water 

 which had lost six parts per million of its salts 

 (mainly magnesium and calcium), 15 c.c. of 

 nitrogen and 2 c.c. of carbondioxide per liter 

 by boiling was about the same as to difference 

 in oxygen content. All the fishes were de- 

 cidedly negative in their reaction to increased 

 carbon dioxide. The differences tried varied 

 from 5 to 60 c.c. per liter above that in which 

 the fish had been kept. When increased car- 

 bon dioxide accompanied low oxygen the neg- 

 ative reaction was very marked; the fishes 

 turned back when the gradient was encoun- 

 tered and only rarely entered the part con- 

 taining the highest carbon dioxide and lowest 

 oxygen. 



Several workers have shovra that carbon di- 

 oxide is very toxic to fish. It appears to be 

 much more so than corresponding differences 

 (24 c.c. per liter) in oxygen content. Fishes 

 turn away when they encounter an increase 

 of as little as 5 c.c. per liter. Since a large 

 amount of dissolved carbon dioxide is com- 

 monly accompanied by a low oxygen content, 

 and other important factors, the carbon diox- 

 ide content of water (strongly alkaline waters 

 excepted) is probably the best single index of 

 the suitability of that water for fishes. The 

 methods . and these results, as well as others, 

 will be published in detail elsewhere as soon 

 as they can be prepared. These aspects of 

 the results are sufficiently different from 

 what workers appear to have been expecting 

 to justify their publication here on account of 



