- 7. The ovary cut transversely, 3-celled.
- 8. The same cut vertically, ovules erect.
- 9. A small but mature fruit, natural size.
- 10. The same cut transversely.
- 11. A portion of a leaf magnified to show the pubescence—with the exceptions mentioned, all more or less magnified.
EXPLANATION OF PLATE 52.
- 1. Dodonaea Burmanniana, (D. C.) natural size.
- 2. A bisexual flower, but with the male organization highly developed.
- 3. Stamens removed to show the ovary.
- 4. The ovary cut transversely.
- 5. The same cut vertically, showing in this instance solitary ovules.
- 6. A bisexual flower, the female organization predominating.
- 7. Stamens.
- 8. The ovary cut transversely.
- 9. Cut vertically, showing the ovules double and superposed, the one ascending the descending.
- 10. Detached ovules, showing their large curunculoid funeculus.
- 11. A different view of the same.
- 12. A mature fruit showing its winged carpels, natural size.
- 13. The same cut transversely, natural size.
- 14. A mature seed—with the exceptions mentioned, all more or less magnified.
XXXVIII.—MILLINGTONIACEAE.
This is a small Indian order limited as yet to a single genus of arborious plants, distinguished by their alternate exstipulate leaves, which are either simple or pinnate, and in the latter case are rendered still more remarkable by their being either abruptly or unequally pinnated. The inflorescence is panicled, the panicles being either terminal or axillary near the summits of the branches. The flowers are small sometimes almost inconspicuous, nearly sessile, on short lateral peduncles of a pale greenish yellow and very numerous.
Sepals 5, persistent, unequal, somewhat in a double series : aestivation imbricative. Petals 5, inserted on the margin of the receptacle, deciduous, alternating with the sepals of two kinds; three outer ones orbicular, entire, with an imbricative aestivation; two interior smaller, acutely bifid, resembling scales. Stamens 5, opposite to the petals, and slightly united to them at the very base ; three exterior sterile, opposite to the larger petals; two interior fertile, opposite to the bifid petals; filaments of the fertile stamens flat : anther, cells globose, dehiscing transversely, placed side by side on the inner side of the saucer-shaped connectivum. Disk, flat, thin, hypogynous, free, except at its point of attachment with the ovary ana receptacle. Ovary ovate, 2-celled; ovules 2 in each cell, superposed. Style simple, short, and thick. Stigma slightly 2 lobed. Fruit a 1-celled, 1-seeded drupe; the dissepiment evanescent above, hardened and persistent at the base. Seed with a small cavity on one side, near the base. Albumen none or extremely thin. Embryo curved : cotyledons thin, foliaceous, folded : radicle curved, pointing to the hilum. Trees. Leaves alternate, without stipules, entire, or rarely pinnated. Inflorescence in panicles, terminal, or axillary near the extremity of the branches. Flowers small, inconspicuous, nearly sessile, on very short peduncles that are arranged along the horizontal branches of the panicles. (W. and A. Prod.)
Affinities. The affinities of this order are viewed in very different lights by different authors. Mr. Arnott and myself aided by a suggestion of Dr. Hooker placed it between Sapindacece and Meliacece. Dr. Lindley looks upon the order as forming but a section or sub-order of Sapindacece, while Meisner for reasons which do not appear refers it to Berberedece, a distribution in which 1 cannot coincide. In confirmation of Dr. Lindley's view it may be stated that my Millingtonia Arnottiana is actually the plant we have described as Sapindus microcarpus, which may be considered a convincing proof of the close relationship existing between them, if not a satisfactory indication that we were premature in separating this genus as the type of a distinct order. However, bringing analogy to bear on the question, we have already seen Hippocratiacece established on its unsymmetrical flowers, and Erythroxylece separated from Malpighiacece on account of the appendages of its petals, hence, if such reasons are applicable to one set of instances they ought to be equally so to another, and as the differences of the arrangement of the flowers between Sapindaceae and Millingloneaceae are certainly equal in amount to those of the other, it must be admitted that if they are to be adopted in the one set they ought equally to be so in the other. In urging this view I do not advocate its adoption, but think with Lindley that the sooner we can retrace a false step the better, and
