indicates a colder environment than is really normal. As the missing 

 years all had relatively late starting and light ice seasons, it is pre- 

 sumed that the March sea temperatures were warmer than normal these 

 years and not reflected in the normal chart of 1-15 March presented 

 here. The charts for the 16-31 March, 1-15 April, 16-30 April, and 

 1-15 May periods are based on data from each of the 13 years from 

 1951-63 and can be considered true normals for those years. The 16-31 

 May chart and the 1-15 June charts, based on 12 and 11 years data re- 

 spectively out of 13 years cannot be considered true normals, as the 

 missing years were light with presumably abnormal warm environ- 

 ments. The 16-30 June chart, which is based on only 7 years' data, is 

 probably more representative of abnormally cold temperatures, as the 

 available data were for the latest and heaviest 7 ice seasons of the 13 

 years. A July chart was not constructed as July data were available 

 for only 2 years, 1957 and 1959, both abnormally cold years accom- 

 panied by very heavy and late ice seasons. 



Two methods were used in the construction of the normal charts. 

 First, the average temperature was determined for each 1° quadrangle 

 for each bimonthly period, and the constant temperature lines were 

 drawn up accordingly. While it was expected that this method would 

 suffice for most of the area of responsibility (39° N., 49° N. and 42° W., 

 60° W.) it was suspected that the method might not be sufficiently 

 discriminating along the east slope of the Grand Banks and at the 

 Tail of the Banks especially in the early Spring. The reason for this 

 suspicion was the fact that the Labrador main stream is quite narrow 

 along the east slope and at the Tail, narrowing in places to less than 

 25 miles in width especially near 45° N. Additionally, the Labrador 

 Current is well defined at the surface during the first half of the ice 

 season. As considerable warming of surface waters takes place be- 

 ginning in May, the Labrador Current becomes less obvious from sur- 

 face temperatures. Therefore the graphical method was also used to 

 more accurately construct the normal isopleth for each normal chart. 

 For example, each available 32° F. isopleth for a given calendar period 

 was traced and the normal 32° F. isopleth was then graphically deter- 

 mined. The same was done for the 36° F. isopleth and the others. It 

 was discovered that the method of construction of normal isopleths by 

 using average temperatures in the 1° of latitude and longitude quad- 

 rangle did indeed fall dow^n in the areas suspected. In other areas both 

 methods correlated very well. The resultant normal charts presented 

 are based on a combination of both methods and are reasonably accu- 

 rate if the limitations previously discussed are taken into account. 



An analysis of the normal charts constructed yields the following 

 conclusions : 



1. The coldest surface environment on the Grand Banks occurs 

 during the last half of March. A comparison of the 15-28 February, 



36 



