We are told, again, that there may be a difference in the number of rings on the two sides of a tree, even at the same level, in other words, a given ring may not be everywhere of the same thickness. This is noticeable in the specimen, Fig. 69. In the famous yew of Darley Dale, Derbyshire, it is asserted that the number of rings varied from 33 to 66 in an inch of radius, taken horizontally--a curiously neat ratio[957]. Michel Montaigne, so far back as 1581, pointed out that the rings were narrower on the north side of the tree. Yet these occurrences do not hopelessly affect our conclusions. The writer possesses a section of a branch of yew from Offchurch, Warwickshire, which was over-developed on one side, the result of proximity to a stream, and of a sunny aspect. The consequent curvature of the heart wood, though certainly disappointing to the bowyer, who bought the tree, did not prove very troublesome even to the amateur ring-counter. Taking the average of a long series of years, and examining a large number of specimens, the inequalities in such specimens would be found to cancel each other. In an aged tree like that of Darley Dale, the coalescence of the rings would be far more perplexing than the unsymmetrical growth. With respect to one source of error, Dr Lowe seems to answer his own objection. “A tree may have died on one side,” says he, “or may have ceased forming, while the other side is growing vigorously.” Yes, but in that case the error would be one of under-calculation; the yew would be credited with fewer years than the measurements warranted. Once again, it is argued that not only may a particular ring have inequalities of width, but the different rings vary among themselves in thickness. And we have seen that growth does not increase uniformly with age. An oak of 50 years had a circumference equal to another which was four times that age. Nor is the rate of growth always uniformly diminished as the tree becomes older. De Candolle found an oak, 333 years old, which showed as great an increase between the rings of 320 and 330 as between those of 90 and 100 years. Now it is not mere captiousness to remark that these figures refer to an oak, not to a steady-growing yew, though Dr Lowe claims that the observation would apply equally well to the latter tree[958]. We may repeat: while variations in rainfall, temperature, and food-supply, correspondingly affect the rate of growth, it is a matter of common knowledge that our seasons tend to follow ill-defined cycles, and that a series of such cycles may be expected to equalize each other with a fair approximation to exactitude. [Illustration: FIG. 71. Yew at West end of Tandridge churchyard, Surrey. Though hollow, it is one of the finest specimens in England.]