532 R. A. SPAETH 



tation. They were brought, three or four at a time, from the 

 aquaria into the laboratory where they were kept temporarily in 

 a shallow white enamelled dish (6 by 26 cms.)- In this, after a 

 brief interval, they assumed the light color phase which served 

 as a uniform starting point. To avoid the carrying over of sea- 

 water into the salt solutions the fishes were always rinsed in dis- 

 tilled water before removing the scales. 



The salts used throughout these experiments were of the high- 

 est quality (Merk, Kahlbaum, and Baker and Adams). Volume 

 molecular solutions (molecular weight in grams up to 1 litre) 

 were made up which were afterwards diluted to the required 

 concentrations with glass distilled water. In a few cases, po- 

 tassium iodide for example, in which the solutions are unstable 

 at molecular concentration, the solutions were made up as needed 

 from the dry salts. 



The majority of the experiments were carried on in Syracuse 

 watch-glasses which were kept covered to prevent evaporation 

 except while under actual observation. 



When the actions of several salt solutions were to be compared 

 it became necessary to immerse all the scales simultaneously. 

 This was accomplished by a simple device consisting of glass 

 rods inserted into a large cork. The free ends had been prev- 

 iously heated and flattened. The scales were placed on these 

 flattened ends — there being a rod for each solution. By dip- 

 ping the ends of the rods simultaneously into the solutions and 

 gently shaking, the scales easily floated off. They were then 

 examined and in case they had been turned face down in this 

 process they were reversed. In this way as many as twenty- 

 five scales (five rods, each carrying five scales), could be manip- 

 ulated at the same time. 



Details of the apparatus used in the oxj^gen and temperature 

 trials are given in connection with those experiments. 



The work has been done chiefly in the Zoological Laboratory 

 of Harvard University under the supervision of Prof. G. H. 

 Parker. This opportunity is gladly taken to acknowledge my 

 indebtedness to him for much kindly criticism and suggestion. 



