CAXADIAX FISHERIES EXPEDITION, lUl'rlo 



95 



narrow, grew broader and broader as the summer progressed, until September, when 

 a period of stagnation set in. Supplementary investigations have shown that at somo 

 time or another during the months of ^farch or April, small herring are found, some 

 of which have only broad summer zones (one or two),, while others have either a broad 

 inner zone and a very narrow outer one, or two broad ones innermost, and one very 

 narrow beyond {vide plate VII, fig. 19). Fig. 20 shows schematically the manner in 

 which the annual rigs were seen to appear at different times of the year. Observa- 

 tions extending over several years have shown that the small herring taken near 

 Bergen only exhibit this narrow outer summer zone in the spring, and that the outer 

 summer zone is in autumn invariably found to be broad. 



Continuous observations of this nature will be analogous to the observation of 

 herring kept in tanks, with periodic examination of the scales. If an experiment of 



December 1909 



UUimo April 1910 



Ultimo May 



Ultimo June 



Ultimo July 



Ultimo August 



Medio September 



Medio October 



Fig. 20. 



this kind, with herrings in tanks, were found to give the same results, we should then 

 certainly be justified in concluding that the summer zones of the scales were formed 

 and developed during the period, April to September, and that the winter rings in 

 consequence represent the time from October to March. In the present instance, this 

 absolute certainty is not attained, since the observations with captured herring 

 naturally to do not represent different states of the same fish at different times. 



During a period of years, from 1907 to 1916, samples of so-called spring herring 

 were collected, these being the fish which early in the spring move in towards the west 

 coast of Norway to spawn. Scales of each fish in the samples collected were examined, 

 and the herring grouped according to the number of rings on the scales. It was then 

 found that for the greater part of the i)eriod embraced, a remarkable regularity 

 prevailed, as indicated by fig. 21. This has been lent from Hjort (V) and 

 shows the percentage of fish (in the samples investigated) falling to each group accord- 

 ing to the number of rings. It will be seen that in 1908, there were many with 4 

 rings ; in 1909, 5 ; in 1910, 6 ; and so on until 1914. The investigations of 1915 and 

 1916 revealed the presence of a large number of herring with eleven and twelve rings, 

 respectively. This regularity, which, by the way, is also encountered in samples of 



