FUNDULUS AND THE SURROUNDING MEDIUM. I2/ 



average of the other five, all dead at the end of twenty hours, 

 was an increase in weight of 19.0 per cent. In a third experiment 

 it was clear that an increase in weight was greater in some speci- 

 mens than in others. While all were dead at the end of twenty- 

 three hours, yet three of the nine experimented with were dead 

 at the end of nine hours. Three others appeared "sick"; they 

 would swim on the side, turn over, follow the bottom, then start 

 to swim again. The gain in weight in these was above that of 

 the others still alive. In other words the results of these three 

 experiments indicated that where there was an excessive gain in 

 weight this was followed by early death. A comparison with 

 these results with those of a fourth experiment about to be 

 described in detail demonstrates the difference in different lots 

 of fishes which in external appearance seemed normal. In this 

 fourth experiment lasting over two days during which eleven 

 weight determinations were made of nearly every specimen this 

 differentiation was all the more striking and because of this, the 

 experiment will be described in more detail. 



The actual weights are not given. The table gives the per- 

 centage change in weight at each period which is found by com- 

 paring the weight of each specimen at the end of the period in 

 question with the original loss in weight. The sign " + " means 

 a gain in weight, while the sign " ' means a loss in weight. 

 The results of this experiment are shown in Table I. The first 

 noticeable feature in this record is the individual variation in the 

 change in weight in all of the specimens. At the end of the 

 observational period seven are dead and eleven are alive; while 

 one is recorded as being lost. All of those alive at the end of 

 the experiment weigh less than they did at some period after 

 the beginning of the sojourn in fresh water. Indeed five weigh 

 less than they did at the beginning of the experiment. If the 

 changes in individual weight be followed, it will be seen to be a 

 case of ups and downs in weight increases. All the specimens 

 show r a gain in weight during the first period. In some the gains 

 are greater than in others. Specimen 5 with an initial gain of 

 8.0 per cent, recovers from this and later loses in weight and is 

 alive at the end of the period. Specimen 18, on the other hand, 

 shows a gain of 2.1 per cent, at first but this increases steadily 



