CANADTAy FisIlERIEfi EXPEDITION, 191 ',-15 167 



If we take the sample of October 4 and plot a curve to show the frequencies <if the 

 various length groups (interrupted line in fig. 1) we see very definitely a bimodal 

 condition with two age groups represented, for one of which the length of 12 cni. may 

 be taken as representative and for the other 19 cm. There is, however, a decided 

 diflFerence in the ranges of the two groups. The smaller one may be considered to have 

 a range of 4 cm. (11-14) and the larger of 8 cm. (16-23). This might be due to the 

 phenomenon of dispersion, the older group showing a broad low curve, and the 

 younger group a narrow high one. I do not believe that this is the full exjjlanation. 

 The range is too great in the older group. It probably indicates that the older group 

 is only apparently homogeneous, that it really consists of two age groups so similar in 

 size as to fuse and give a good unimodal curve. Other considerations to be mentioned 

 later support this view. 



The sample of September 28 shows a similar condition. The practical ranges of 

 the two groups would be 3 and 5 cm., respectively. The significance of this would seem 

 to be that by the third year the spring and fall spawned schools have fused into a single 

 group. 



The third sample (III) of September 16, from Bocabec shows imperfectly a tri- 

 modal curve (fig. 1, dotted line). The sizes representative of the three groups may be 

 taken as 12, 15 and 19 cm. The ranges are 3 cm. (11-13), 3 cm. (14-16) and 5 cm., 

 respectively. The first and third of these groups are evidently identical with the two 

 groups of the sample of October 4. The second group (15 cm.) was doubtless present 

 in the latter sample but not in sufficient numbers to appear distinctly. 



Let us designate these three groups A (19 em.), B (15 cm.) and C (12 cm.). 

 B and C give a bimodal curve with a total range of 6 cm. The growth of the smaller 

 group (C) appears to continue farther into the fall than that of the larger group (B). 

 This would bring them close together and make them fuse into one group vrith a 

 range of 5 cm. and a mean size of 14 cm. as seems to be the case in the samples of 

 November 3 and November 10. (for the latter see the curve in fig. 1 with alternate 

 dot and dash). In this latter sample the larger group with a mean size of 19 cm. is 

 evidently A and the smaller group with a mean size of 14 cm. represents (if our inter- 

 pretation be correct) B and C fused. In the spring of the year group A seems to have 

 been in the same condition as shown in the sample of April 10, with a range of 5 cm. 

 and a mean size of 14 cm. 



The degree of fusion of B and C and the relative abundance of the two groups in 

 the various samples give a varying picture as shown in the samples of October, 

 November, and December from St. Andrews. 



In the middle of September samples from widely separated localities along the 

 coast were examined and also a number of samples from the same locality in order to 

 determine whether the mean size of an age group varied greatly in the different locali- 

 ties and in different samples from the same locality. These samples were in great 

 part obtained through the courtesy of Captain Calder of the Seacoast Canning Co., 

 Eastport. The localities were Jonesport (Maine), Grand Harboiir (Grand Manan), 

 Lepreau, Pocologan, L'Etang, and Bocabec. The samples sht»wed uniformly a great 

 preponderance of the A group. The mean size varied, being IT cm. (Jonesport, 

 Lepreau, and Pocologan), 18 cm. (L'Etang), and 19 cm. (Grand Harbour and Boca- 

 bec). Evidently there is an appreciable difference in the size of the same age gi'oup 

 from different localities. 



Samples were taken from several boats bringing herring from Bocabec on 

 September 16. These showed uniformly a preponderance of the A group with in each 

 case a mean size of 19 cm. The same is shown in a sample of September 28 from Oak 

 Bay. This shows that herring from the inner side of Passamaquoddy bay may be 

 considered uniform and treated together. Those from points as far away as L'Etang 

 must be treated separately. The differences shown in the samples of September 16 

 from Bocabec indicate the amount of uncertainty to be associated with deductions 



