IF I have rightly interpreted General Strachey's courteous criticism of my paper on the Carnatic rainfall, the gist of his objections may be summed up by saying that the method by which I endeavoured to estimate numerically the genuineness of the apparent cyclical variation of that rainfall, as a recurrent phenomenon, is logically invalid. This, I must frankly admit, is really the case; my error has been somewhat of the nature of a petitio principii, and is indefensible. I have reasoned upon a series of values directly obtained from the observations, as if they had been obtained deductively from some independent source, and had been found, on trial, to agree, within certain allowable limits, with the results of the observations. This procedure, as General Strachey has shown, is manifestly illogical; and the inferred “high probability that the apparent undecennial fluctuation of the Carnatic rainfall is no chance phenomenon,” in so far as this conclusion depends on the above erroneous reasoning, necessarily falls to the ground. But only in so far. The validity of the data afforded by the registers remains, of course, unaffected; and these data, as they stand, seem to me to furnish evidence of so pronounced a character that it is at least improbable that the apparent fluctuation is fortuitous. The considerations on which I base this opinion are:—