1911] Michael: Chaetognatha of the San Diego Region. 129 



in temperature occurs in the winter months and a maximum in 

 June and July. 



These monthly variations in temperature are evidently due 

 to seasonal influences. Consequently, unless some definite period 

 of the year is selected as a basis for determining the effects of 

 variation in temperature on distribution, our interpretations may 

 be misleading. Because the greatest number of collections were 

 made during June and July I have selected these months for 

 this purpose. 



While during these months the temperature varied irregu- 

 larly from 15? 9 to 22? 3 the number of hauls made in water of 

 known temperature are too few to permit considerations of the 

 effects of small changes in temperature on the number of speci- 

 mens obtained. I have, therefore, segregated the hauls into three 

 groups according to whether the temperature varied from 15?9 

 to 17?5, 17?6 to 19?5, or 19?6 to 21?5. I have not considered the 

 only hauls (1568 and 1591) made in water having temperatures 

 above 21 ?5 because, on the basis of two hauls, it is impossible to 

 eliminate the effects of light on the number of specimens 

 obtained. The following table, therefore, supplies data concern- 

 ing the relation between temperature and abundance of S. 

 hipunctata with respect to the months of June and July. 



TABLE 40 

 Surface Distribution of S. Mpunctata with Eespect to Temperature 



Temperature Number of hauls Number Number of specimens 



in r * N of f ^ > 



Centigrade Total Successful hours Total Per hour Frequency 



1.5?9-17?5 21 19 12.4 6818 549 91 



17?6-19?.5 53 36 42.8 16274 380 68 



19?6-21?5 12 7 6.4 1422 222 58 



This table shows clearly that the average number of speci- 

 mens obtained per hour from the surface is greatest when the 

 temperature of the water is least. It also shows that the fre- 

 quency is greatest when the temperature is least. 



It may be objected that, on the basis of so few hauls, the 

 effect of their uneven distribution with respect to intensity of 

 light might account for the results contained in the above table. 



