In the zone between 50 and 55° and further north, the curve dips abruptly, signifying deterioration of conditions for the formation of sub marine canyons. This is explained by the smaller amplitude of fluctuations of ocean level and a decrease in atmospheric precipitation. The curve also dips abruptly south of the zone of its maximum values, at 20-27° lat., where the probability of the formation of submarine valleys is also low. This substantiates Shepard's hypothesis on the comparative scarcity of submarine valleys at latitudes of about 30°.
Between 5 and 20° the curve has two additional maxima, indicating a recurrence of conditions favorable for the formation of submarine valleys. In this zone the amplitude of the fluctuations of ocean level is considerably smaller than in the middle latitudes, but there is a heavy increase in precipitation which more than compensates for the attenuation of the first factor. In this zone the fall in ocean level will not be very large, but the intensive erosion caused by excessive rainfall enhances the incision of canyons, so that the formation of submarine valleys becomes possible.
Curve C drops again in the equatorial zone, indicating a very low probability for the formation of submarine canyons in spite of the maximum amount of atmospheric precipitation. This is because the amplitude of the fluctuations of the ocean level approaches zero along the equator.
We have described curve C and determined the latitudes providing the most favorable conditions for the formation of submarine canyons. If our approach toward the solution of this problem is correct, then the maximum number of submarine valleys should be concentrated at the latitudes corresponding to the maxima of the curve. We must now check whether the regions providing the most favorable conditions for the formation of submarine canyons indicated by the graph are actually those of their most extensive occurrence in nature.
We must first study the distribution of submarine valleys at different latitudes. We found suitable data for solving this problem in the work of Shepard /11, 12/, Klenova /5/, and Saks /9/. Data on more than 200 submarine valleys taken from these sources are listed in Table 2.
They were used for plotting the graph in Figure 4 illustrating the latitudinal distribution of submarine canyons. Each point of this curve corresponds to a known submarine canyon at this latitude. In order to render Figure 4 more informative we transferred to it also curve C from Figure 3 plotting the variation in the conditions for the formation of submarine valleys.
It is now possible to check, by comparing the two curves, whether the maximum occurrence of submarine canyons is actually found within the regions that should provide the most favorable condition for their formation according to the graph. In Figure 4, the great similarity of the two curves,
both in main features and in details, is immediately apparent. For example, both curves reach their maximum values in the middle latitudes, in the zone 30 to 50°; each reaches a secondary peak in the latitudes 13 to 18°, somewhat lower than the maximum value. In both cases there is a depression between these two peaks in the region of latitudes 25 to 27°. The minimum values of both curves occur on the equator and at latitudes higher than 60 to 65°, and so on.
It is our opinion that the similarity of the two curves cannot be accidental, and that we are justified in assuming that a causal relationship exists between the phenomena illustrated by the graphs. Similarity of the curves shows that the assumed pattern of the formation of submarine valleys is to a certain extent reflected by the actual distribution of canyons. The hypothesis that the genesis of submarine valleys is actually related to fluctuations in ocean level due to deflections of the rotation axis within the Earth evidently reflects, somewhat, the real conditions.
The distribution curve for submarine valleys was plotted from data that we found in scientific literature. Obviously new submarine canyons will be discovered in the course of time and their plotting on the graph may modify it to some extent. Our final judgement on the similarity between the curves of the most favorable conditions for the formation of submarine canyons and their actual distribution must be withheld until curve D has been marked with all submarine canyons actually existing in nature. Nevertheless, our data was sufficiently extensive to reveal the general distribution pattern of submarine valleys. Probably the nature of this pattern will not be significantly affected by supplementary information on new canyons. In order to trace the correlation between the phenomena illustrated in Figure 4 the curves need not resemble each other in all details. If, after the accuracy of the graphs has been improved, the curves retain their similarity with respect to their principal features (such as their maxima in the zones 30 to 50° and 5 to 20° and the well marked depressions between them), we shall be justified in assuming that the phenomena reflected by these curves are in some way correlated and that the conclusions reached in this paper provide some knowledge of the real conditions.