of the Fishenj Board for Scotland. 365 



would at this stage increase the lower to a greater extent. In July the 

 smallest specimens of this, now the second series, increased from 42 to 

 68 mm., an increment of 26 mm., and the largest from 98 to 127 mm., 

 an increment of 29 mm., a slight difference which may show the tendency 

 of those which have begun to lag in the i-ace to maturity to lag still 

 further, while those which have taken the lead increase it. But any 

 such extension of the range will make very little appreciable diff"erence 

 in the mean size; if the lower limit is fixed at 35 mm., the range becomes 

 63 mm., and the average size 66'5 mm. 



On 19th-22nd August, twenty-eight days later, the first series, or 

 brood of the year, collected, comprising 188 specimens, varied in length 

 from 37 to 75 mm., with a mean size of 52"3 mm. (2y^ inches). The 

 apparent mean increment in the interval thus amounted to 11 '6 mm., 

 while the largest specimens had increased in length by 26 mm., or almost 

 1 mm. daily, and since the smallest taken had increased by 5 mm., it 

 would indicate that the proportional number of the very small specimens 

 escaping was less than before. But the occurrence of specimens less 

 than 20mm. later in the year shows that the settlement of the swarms 

 of post-larval dabs had not ceased in August, although doubtless greatlj' 

 reduced in ratio. 



Applying the same method, and again placing the lower limit at 

 18 mm., the amended series would have a range of 57 mm. up to the 

 maximum caught, viz. 75 mm., and the mean size would become 46*5 mm., 

 showing an apparent increment of 13"0mm. in tlie twenty-eight days. 

 This approximates to the apparent mean size, computed from the catch, 

 being 5*8 mm. less, which, in like manner, might be taken as equivalent 

 to the error introduced at this period by the imperfection of the net. 

 The apparent increment in the interval, computed from the actual 

 measurements, was 11 "6 mm., while the increment as amended amounts 

 to 13 mm., or 1*4 mm. more. The apparent growth of the next older 

 series in the same period (with a range from 84 to 137 mm., and 

 numbering 1 10 specimens) was only 1"3 mm., which is probably incorrect 

 and too low. 



There are no later hauls in the same locality with the same net to 

 allow the comparison to be extended further, but it is evident that, as 

 the fishes grow in size, a point must be reached at which practically 

 none escape, and that in the interval between the period when the 

 swarms settle, and this point, the i^atio between the numbers which 

 escape and the numbers captured, will constantly diminish, and the 

 error from the net, or the deviation between the apparent and real mean 

 size will diminish in like proportion. In the above cases the assump- 

 tion that the lower limit is equal in July and August, after an interval 

 of twenty-eight days, is no doubt inexact. The lower limit would be 

 probably more correctly placed at 20 mm., and on that basis the ranges 

 would be red\;ced to 55 mm., and the mean size would be 47*5 mm., or 

 4-8 mm. less than the apparent increase derived from the actual 

 measurements. In May, when the series is a year old, the minimum 

 size was 42 mm., and the range 56 mm., for 84 fishes, and at this time 

 probably very few, if any, smaller fishes escaped capture. 



The fine-meshed net was also used in one haul in Aberdeen Bay 

 on 15th January 1902, and the results may be also particularised. 

 The number of dabs taken was 214, measuring from 28 to 80 mm., 

 the range being 52 mm., and the arithmetical mean 47-8 mm., or 

 Ig inches. This is under the average in the Forth in August, and 

 there are reasons for believing that in this case the apparent average 

 size is less and notfgreater than the true mean size. 



In this haul scaixely any other fi.sh were taken either in the otter- 



