Neither could the precession of the equinoxes have been of itself a principal cause, for here also the limit of time negatives the supposition. This precessional circle carries the perihelion and aphelion, and with it the seasons, completely round, and brings them back to the old position, in about 21,000 years, and therefore if glacial periods were occasioned by them, there ought to be alternations from _maximum_ of cold to _maximum_ of warmth in each hemisphere every 10,500 years. But this has certainly not been the case even in recent times, and still less if we go back to the quaternary, tertiary, and earlier geological periods. In fact it is only when combined with periods of high eccentricity of the earth's orbit, according to Croll's theory, that precession can pretend to have any claim to be an important factor in the production of glacial periods. And even then the question is not of its being the sole or principal cause, but only whether it has had such a perceptible auxiliary effect on other more powerful causes, as may enable us to use it as a chronometer in assigning approximate dates for some of the more important phenomena of the long and varied period between the close of the Tertiary and the establishment of the Recent period. As man certainly existed throughout the whole of this period, the possibility of finding such a chronometer becomes intensely interesting, and I proceed to discuss the latest state of scientific opinion respecting it. But as Croll's theory if a real is clearly only an auxiliary cause, I will, in the first instance, point out what are the certain and admitted causes which account for variations of temperature irrespective of latitude.