But the obstacle which most effectually blocks the use of Onderdonk's classification in the systematic arrangement of peaches is the brood of hybrid seedling peaches annually brought forth by fruit-growers. No doubt the classification is workable, to a degree, with the type-varieties and a few carefully selected progeny but after the practical peach-grower, with a devil-may-care attitude toward classification, crosses and recrosses the types, the several races become hopelessly interlocked. The characters chiefly used by Onderdonk, as has been said, are fluctuating variations and these do not descend according to Mendelian laws. And so the great out-pouring of varieties during the past quarter-century has literally swamped a classification which served only fairly well when it included but the pioneer varieties. In the trituration of the thousand and more varieties of peaches now going on, the Onderdonk classification will be less and less useful. In dismissing the Onderdonk scheme as having but limited application for classificatory purposes, acknowledgment is made that it serves other purposes very well. It calls attention to the history of the peach; it shows that racial strains of the peach are arising; it brings out valuable information in regard to hardiness and the rest-period of peaches; it offers instances of modification of the peach by climate; and it shows the capacity of the peach to vary. For thus illuminating the natural history of the peach, more especially the climatology of the peach, pomology is much indebted to Onderdonk and Price. _A key to varieties of peaches._--A natural classification of peaches to show the relationships of varieties is seemingly impossible. The deluge of new varieties, which growers continue with cheerful optimism to pour out, overwhelms the classifier with difficulties. About the best that can be done is to arrange varieties, for convenience in identifying, according to some of the artificial systems of a century ago when the cult of the classifier was at its height. These were really synoptical keys rather than biological classifications. If such a key is to be used very generally by fruit-growers, only characters of the fruit are admissible, thereby attaining necessary simplicity and providing that all data can be had at one examination.