A thin layer of sand was added in the respiratory chambers to 

 meet the thigmotactic requirements of the shrimp. On smooth bot- 

 tomed surfaces shrimp do not rest but tend to swim almost constantly. 

 Addition of sand naturally reduced their activity to normal. Test 

 animals were introduced into eleven chambers, keeping the twelfth 

 chamber empty as a control. Feeding was suspended from one day prior 

 to the experiments through the rest of the experimental period. 



The test animals were allowed to become accustomed to the res- 

 piratory chambers for about 12 to 16 hours. During this period they 

 were maintained in conditions similar to those of their control 

 and received seawater from reservoir 4. After the initial adjust- 

 ment period the animals were gradually exposed to the test media 

 supplied from reservoir 1. The time taken for replacing the 

 control medium by the test solution varied depending on the test 

 salinity and the temperature change. Salinity changes took place 

 in 15 to 60 minutes. Temperature changes from 25° to 18° or 32°C 

 occurred in less than two hours. However, extreme temperature 

 changes from 18° to 32°C or vice versa required up to three hours. 

 The first hour readings were taken as soon as the test salinity and 

 temperature conditions appeared in the respiratory chambers. Zero 

 hour readings at the three different test temperatures represent 

 the oxygen consumption in the respective control conditions of 

 15°/oo and 25°C or 18° or 32°C. 



Respiratory rates (ml O2/L) were determined on the basis of 

 differences in dissolved oxygen levels recorded between the animal 

 chambers and the reference chamber. For the sake of comparison of 

 the responses exhibited by shrimp of different weights, the res- 

 piratory rates were converted to ml 02/L/g. The oxygen levels were 

 analyzed with a galvanic cell oxygen analyzer calibrated against the 

 Winkler method. It should be added, however, that the readings ob- 

 tained with the oxygen analyzer at 18° and 32°C were not reproducible 

 probably due to the fact that the instrument was not temperature 



45 



