PLANKTON OF THE GULF OF MAINE 
247 
is evident that M. longa fluctuates widely in the gulf from year to year, being ex- 
tremely rare, if not altogether absent, in some years but widespread in others. The 
years 1912 and 1914 and the summer of 1916 were periods of scarcity, while 1915, 
the winter of 1916-17, and 1920 were times of plenty. The relationship of tempera- 
ture to these annual differences is discussed below (p. 252). 
Seasonal distribution . — During the years 1915, 1920, and 1921, which may be 
taken as representative of the periods when M. longa is at a maximum in the gulf, 
it was taken at the following percentages of the stations : 
Months 
Percentage 
of stations 
Months 
Percentage 
of stations 
100 
60 
February 
17 
August 
75 
March 
74 
September 
60 
April 
87 
86 
May 
72 
87 
This suggests that on the whole M. longa is apt to be found most widespread 
in the gulf during the late autumn, winter, and early spring, and least so during the 
summer and early autumn. The low percentage of stations at which it was recog- 
nized in February, 1920 (only station 20046), would upset this rule were it a regular 
annual phenomenon; but it is more likely that that month marked the beginning 
of a period of abundance which endured throughout 1920, and that still fewer stations, 
if any, would have yielded it during the preceding January or December. In fact, 
a February station was most prolific of this species at St. Andrews during the winter 
of 1916-17, as noted above (Willey, 1921). 
Seasonal fluctuations in the actual abundance of M. longa, as reflected in the 
numbers of specimens per square meter, did not parallel the seasonal rise and fall in 
the percentage of stations at which it occurred, it being much more plentiful in the 
vertical hauls in August and October than from March to June or in September of 
the years 1915 and 1920, as shown in the following table: 
Date 
Average 
number 
per square 
meter at 
stations 
where it 
occurred 
Average 
number 
per square 
meter, all 
stations 
included 
Date 
Average 
number 
per square 
meter at 
stations 
where it 
occurred 
Average 
number 
per square 
meter, all 
stations 
included 
March, 1920 
990 
1, 650 
2, 504 
3, 193 
692 
1, 429 
1,808 
1,552 
14, 850 
2,453 
8,601 
13, 637 
1, 533 
7, 280 
April, 1920 
May (1915 and 1920 combined) 
June, 1915 
October, 1915 
It is unfortunate that only four vertical hauls were made during August, 1915, 
when the species averaged so much more plentiful than we have ever found it before 
or since in the gulf. It may have been only a chance that the net hit local swarms, 
and more vertical hauls might have proved barren of M. longa , thus reducing the 
month’s average. However, the fact that this northern species should have been 
so plentiful (from 10,300 to 23,400 per square meter) at three late summer stations 
