In many regions seed-propagation is the only method which has been used with this plant. In the United States, in Madeira, in Algeria, and in the Philippines, cherimoyas have been grafted and budded successfully; one or the other of these methods should be employed to perpetuate choice varieties.

If kept dry the seeds will retain their viability several years. Given warm weather or planted under glass, they will germinate in a few weeks. Under glass they may be sown at any time of the year; if in open ground, they should be planted only in the warm season. Seeds should be sown in flats of light porous soil containing an abundance of humus, and should be covered to a depth of not more than 3/4 inch. When the young plants are three or four inches high, they may be transferred into three-inch pots. Good drainage must be provided, and they should not be watered too copiously. When eight inches high they may be shifted into larger pots, or set out in the open ground. In the latter case, they must have careful attention, and, preferably, shade, until they have become well established.

For stock-plants on which to bud or graft the cherimoya, several species of Annona have been employed. A. reticulata, A. glabra, and A. squamosa are all recommended by P. J. Wester. In Florida A. squamosa has proved to be a good stock when a dwarf tree is desired; A. glabra tends to outgrow the cion. In California, seedling cherimoyas as stock-plants have given the best results.

Shield-budding has worked very satisfactorily in the United States. In several other regions horticulturists have found grafting more successful. Budding is best done at the beginning of the growing season, when the sap is flowing freely. Stock-plants should be 3/8 to 1/2 inch in diameter. Well-matured budwood from which the leaves have dropped is preferable, and it should be gray, not green, in color. The buds should be cut l 1/2 inches in length, and should be inserted exactly as in budding avocados or mangos. Waxed tape, raffia, and soft cotton string have proved satisfactory for tying. Three or four weeks after insertion of bud, the wrapping should be loosened and the stock lopped at a point 5 or 6 inches above the bud. Wrapping should not be removed entirely until the bud has made a growth of several inches.

For grafting, two-year-old seedlings are to be preferred (for budding they may be somewhat younger). The cleft-graft is the method usually employed. The cion should be well-matured wood from which the leaves have dropped. C. H. Gable wrote from Madeira in 1914 : "I have been surprised to find how easily the annona is grafted. My first few efforts were not very successful, but later I grafted them in all sizes from seedlings smaller than a lead pencil to old trees, and more than 90% have grown beautifully." Gable found it advisable after making the graft to paint the cion and the top of the stock (around the cleft) with melted wax, to prevent evaporation.

Old seedling trees can be top-worked without difficulty. For this purpose cleft-grafting is used more commonly than any other method.

The pollination of the cherimoya has been investigated in Florida by P. J. Wester, and in Madeira by C. H. Gable. It has been thought that the scanty productiveness of many trees might be due to insufficient pollination, and the investigations tend to confirm this belief. Gable reports that normally in Madeira not more than 5 per cent of the flowers produced develop into fruits. By hand-pollinating them, however, he was able to obtain thirty-six fruits from forty-five flowers.

After carrying on pollination experiments in Florida during several years, P. J. Wester 1 wrote: "The investigations indicate that the flowers of the cherimoya, the sugar-apple, the custard-apple and the pond-apple are proterogynous and entomophilous, though the pollinating agent of the last-named species has not been detected." A proterogynous plant, it may be remarked, is one in which the pistils are receptive before the anthers have developed ripe pollen, cross-pollination being therefore necessary, and some outside agency being required to effect it. In the case of the annonas the work is done by insects; hence the plants are termed entomophilous.

The pollination of the closely allied Asimina triloba is thus described by Delpino: 2 "The stamens project in the center of the pendulous protogynous (proterogynous) flower as a hemispherical mass, from the middle of which a few styles with their stigmas project. In the first (female) stage of anthesis the three inner petals lie so close to the stamens that insect visitors (flies) cannot suck the nectar secreted at the bases of the former without touching the already mature stigmas. In the second (male) stage the stigmas have dried up and the inner petals have raised themselves, so that the anthers, - now covered with pollen, - are touched by insects on their way to the nectar. Cross-pollination of the younger flowers is therefore effected by transference from the older ones.,, Wester concluded that one cause of the unproductiveness of the cherimoya in Florida was the scarcity of pollinating insects. Even under the same conditions of environment, however, there are marked differences in productiveness among seedling trees. The subject deserves further investigation. Productive varieties especially should be studied, to determine whether or not they differ in any way from the typical less fecund form in manner of pollination.

1 Bull. of the Torrey Bot. Club, 37, 1910.

2 Paul Knuth, Handbook of Flower Pollination.