53 
Any one of the above may show the desired, correlation with bird data 
but» the selection of which, or which combinations, will produce the i..o-^ 
interpretable correlation appears to be variable. Dr. Robert Pyle found 
in the analysis of November 190o data (biC—Ibj, thao one 150—meter oc...— 
oerature showed the best correlation with bird numoers and uiiuo surface 
temperatures alone did not. (l) March data (3)? however, showed poor 
corelation with 150-meter temperature but good correlation with 100- 
meter temperature (3A) and fair correlation with surface temperature 
(3B). The March data also showed good correlation with depth of the 
mixed layer (3C). April correlations with surface and 100-meter tempera¬ 
tures were good (4), and 150-meter temperatures were interpretable but 
not as nice as the November structure. In January, an excellent measure 
of at least the 5° N. area Vas the first derivative (slope) of the 
thermocline depth curve (2A). January surface temperatures easily 
located the 12° N. area (2B), and although the surface temperatures are 
not at hand for all of the 5° N. region (2C) they also probably showed 
good correlation. 
A consistently good correlation exists between' sun-ace temperature 
and the avifaunal zones at 5° N. and 12° N. The January surface tem¬ 
perature shows a rise and fall of almost 2° C. between 11° N. and 13° N. 
(23). The March picture (3B) for the same latitude is similar but 
slightly broader, ca. 11° N. to l4° N. April (4 b) finds the typical tem¬ 
perature rise and fall present but the range is slightly less than 1° C. 
and is generally warmer than either the January or March temperatures. 
The width of the zone is smaller and displaced southward 10° N.-ll° N. 
The features of the three sets of data are plotted together for compari¬ 
son (see 5A). Note the tendencies toward l) Increasing temperature; 
2) "Flattening" of the temperature range; 3) Equator-ward shifting in 
location. The specific oceanographic processes involved in this are not 
known to me but a logical assumption is to presume that these changes are 
results of differential heating of the current generating waters involved, 
and, likely, the trends noted are seasonal. 
Good thermal data for the 5° N. region is at hand only lor the April 
section, but special attention was given to the region at that time. Con¬ 
current observations from the three survey vessels, more Irequer.t 3T 
casts, and backtracking to fill in nocturnal gaps in observation were 
carried out for all or part of three days in the area of bird concentra¬ 
tion. The nominal data for 22-28 April are plotted in Graph 4. Since 
both surface-and deep-water data have proven to be useful for correlating 
vith bird abundance two types of BT iniormation are plowed. j.i.e s*^r^ace 
section includes the sea surface temperature and the 100-meter temperature. 
The 100-meter temperature was chosen because l) it had had a good cor¬ 
relation in the 12° N. region, and 2) for the most part 100 meters lies in 
the isothermal layer or at least above the thermociine. At any ra^e, uhese 
two measurements move in the same directions and appear to be measuring 
an associated phenomenon. Because two heads are better than one (ii Dotn 
think alike) the two curves are added together, graphically, giving the 
composite curve in Figure 6. And in a like manner the deep-water data 
were graphically added together. 
/ 
/ 
I 
