195 



We do not consider the keeled scales (A;,) to have any 

 particular significance and they seem to show very little 

 variation. For this reason we have not investigated them 



mathematically. 



Our indebtedness to Professor Johnstone is acknowledged. 

 He drew up this scheme of investigation and supervised its 

 carrying out. Thanks are also due to Mr. Smith and 

 Mr. Fleming, who have helped in the measuring and examin- 

 ation of alUhe samples. It should not be overlooked that the 

 measuring and examination alone of a sample of 50 fish is a full 

 day's work for two operators, and the greatest care must be 

 exercised. Scale reading takes up a great deal of time and 

 becomes very trying. The results of scale readings of the 

 samples will be the subject of another report. 



The analysis of all samples received from the commence- 

 ment of these biometric investigations in 1913 to the present 

 time are appended. See Tables XII to XX. 



" Range " of a character in the Tables means the per- 

 centage of the measurement T.cd. occupied by the measurement 

 of the" character under consideration. The material is divided 

 into sub-samples, which appear under the heading of locality, 

 month and year. The final column for each year is the total of 

 these sub-samples and forms the " lumped " samples data. 



Table XX. Character K,. The " range " of these data 

 are not percentages of T.cd. but absolute numbers of the 

 keeled scales as counted. 



