But Croll's theory brings in the secular variation of the eccentricity, and contends that although precession may have little or no effect while the earth's orbit is nearly circular, as it is now, it must have a considerable effect when the orbit flattens out, so that the distances from the sun and the durations of summer and winter become exaggerated. Croll calculated the periods when such _maxima_ and _minima_ of eccentricity occurred for several revolutions back from the formula of the great astronomer Leverrier, and found that going back for the last 260,000 years there had been two _maxima_ of high eccentricity, one 100,000 years, and the other, and more intense, 210,000 years ago, with corresponding _minima_ of low eccentricity between, which corresponded remarkably well with the refrigeration commencing in the Pliocene, culminating towards its close or in the early Quaternary, subsiding into a long inter-glacial period, rising again in the later Quaternary to a second glacial _maximum_ a little less intense than the first, and finally gradually subsiding into the low eccentricity and temperate climates of more recent times; especially as the geological evidence shows many minor oscillations of heat and cold and advances and retreats of glaciers during each phase of these periods, such as must have occurred from the shorter recurrent effects of precession according to Croll's theory. Croll's calculations show that, at the period of _maximum_ eccentricity 210,000 years ago, the earth would have been in mid-winter 8,736,420 miles further from the sun than it is now, and the winter half of the year nearly twenty-eight days longer than the summer half, instead of being six days shorter as at present. It appears, moreover, from a volume just published, _On the Astronomical Causes of an Ice Age_, by Sir R. Ball, one of the highest authorities on mathematics and astronomy, that Croll had understated his case. Ball says that "Croll, misled by a statement of Herschell's, had assumed the number of units of heat received from the sun, in a hemisphere of the earth, as equal in summer and winter. But in reality, of 100 such units, 63 are received in summer and only 37 in winter. As the maximum of eccentricity which is possible would produce an inequality between summer and winter of 33 days, they had the following possible conditions in a hemisphere--summer 199 days and winter 166 days, or summer 166 days and winter 199 days. In each case it must be borne in mind that 63 heat units arrived in summer and 37 in winter. If the summer were a long one and the winter short, then the allotment of heat between the two seasons would be fairly adjusted. The 63 units were distributed over 199 days and the 37 units over 166 days, and a general inter-glacial state was the result on the hemisphere. If, however, a torrent of heat represented by 63 units was poured in during a brief summer of 166 days, whilst,