324 
BULLETIN OF THE BUREAU OF FISHERIES 
Canadian waters on the surface in 31 to 32 per mille when a few fathoms sinking 
would have carried it into much more saline water. 
From the data just outlined it would appear that the whole column of water in 
the offshore parts of the Gulf of Maine offers an environment favorable for the exist- 
ence if not for the reproduction of S. serratodentata during the season (July to Sep- 
tember) when it is most widespread there, but probably it could not long survive 
water much less saline than about 31 per mille or colder than 6° to 8°, and Huntsman 
(1919) has suggested that low salinity may be the factor that bars it from the Gulf 
of St. Lawrence. 
Neither temperature nor salinity offers an explanation for the disappearance 
of S. serratodentata from the gulf in autumn, fcr the water is considerably warmer 
in November than when it first enters the gulf in spring, and the salinity is not very 
different from that of late summer. Neither does its immigration into the gulf in 
spring parallel the vernal warming of the water, but is not at its height until long after 
the gulf is warm enough for its support. It is therefore likely that the increase in its 
numbers with the summer chiefly mirrors an accumulation of the stock within the gulf, 
where it finds good feeding ground and conditions favorable for growth and prolonged 
existence. Apparently no more enter after early autumn, a phenomenon probably 
connected with the seasonal reproductive cycle of the species, and as the visitors of 
summer die off during the autumn from one cause or another or are devoured by 
other animals without leaving progeny to take their places, S. serratodentata disap- 
pears from the gulf, not to reappear there until with the earliest immigration of the 
succeeding spring. 
Our data do not allow a statement as to the vertical distribution of S. serratodentata 
in the Gulf of Maine more definite than that it has seldom been detected there at the 
surface, though most often in hauls from shoaler than 100 meters. If it is actually 
as uncommon right at the top of the water in the gulf as now appears to be the case, 
the food supply may be as effective a factor as any of the physical features of its 
surroundings in holding so rapacious an animal at lower levels. 
There is no evidence that this chsetognath ever succeeds in reproducing itself in 
the gulf. 
Sagitta maxima Conant 
In a previous chapter (p. 64) I have discussed the geographical distribution of 
this species and of the next within the gulf from the standpoint of their routes of 
entrance and dispersal. What demands chief emphasis here is that both S. maxima 
and S. lyra are distinctly seasonal in the inner parts of the gulf, like S. serratodentata. 
During all our cruises we have found only a single specimen of S. maxima within the 
offshore banks during the summer or early autumn months (eastern basin, September 
2, 1915, station 10310), our failure to find it there in July and August, 1914, being 
specially significant because it occurred then off the seaward slope of Georges Bank 
(station 10220). Neither have we any early winter records for it in the gulf; this, 
however, may be an accident, for we have tried only two tows in the deep trough in 
December or January, which may simply have missed the S. maxima. However, 
tins large chsetognath was detected at 12 stations within the gulf as well as over the 
deeper parts of the continental shelf off southern Nova Scotia during March, April, 
