WHEN the author died in 1954, he had nearly completed his work and only a few trivial emendations and additions have been made. All the descriptions were complete with the exception of the section Appendiculati which has been prepared from his account of the series and subseries. In the analytical key there was no method of arriving at the sections, subsections, series and subseries. As it seemed undesirable that the text should have to be consulted to arrive at the appropriate group, the group descriptions have been extracted from the text and incorporated into the key. Under the nomenclature only dates had been quoted, but it seemed desirable to have full references. This has been done with the exception of the group headings. No authority or reference for the group headings, many of which are new or have new status, had been given by the author, and it has only proved practicable to give basonym authorities. All new names and combinations have been published separately in a paper in Watsonia. All the type material from Watson's herbarium has been deposited at the British Museum (Nat. Hist.).

It is believed that the author would have wished to thank Miss Ruth M. Ball and Mr A. W. Darnell for drawing the illustrations; the authorities of the British Museum (Nat. Hist.) and the South London Botanical Institute for the use of specimens and books; and the many fellow botanists for help received during his life-long study of the genus Rubus, which has resulted in this, his most considerable work.

It is the duty of the exploring botanist to expound, not to confound, nature: to bring to light what he meets with, not to impose his own ideas about what ought to be there. Consequently he should not shrink from accepting the fact that, despite heredity, variations occur and lead to the production of new types.

The criteria of a species in Rubus are that it possesses constant and inheritable characters running all through the plant and distinguishing it from all other species. Alongside this assemblage of characters, species are liable to vary from one generation to another in certain characters of a minor kind that are often found also in related species. This faculty of variation does not invalidate the conception of a true breeding species. On the contrary, it is provided for and ensured in the mechanism of reproduction; it may be regarded as an asset, not a disability to the species. As in other groups the species of Rubus differ in size, that is in degree of distinctness; they may be different in mode of origin and in mode of reproduction. They may differ again in their behaviour; they may be able to cross with certain species but not with others; the cross may be sterile, or it may be fertile from the first, or, sterile at first, it may in time become fertile, or even very fertile.

Because of the fluctuating variations referred to, which are controlled to some extent by genes and are liable to be altered in any generation, the characters of the species may not all be present together in every individual; a few will be missing here or there, perhaps occasionally in forms occupying geographically separated areas. If, then, an ice-sheet or an invading sea blots out a part of the population of that species, the parts that remain may not be able to breed together as one unit and may develop along rather different lines. It is possible that a few geographical races may exist of some Rubus species in Wales and south Devon as compared with Sussex and Kent or the west of France.

Genus RUBUS L.

(Type R. ulmifolius Schott f.)

i. Stem herbaceous, rootstock creeping and suckering. Stipules broad, cauline.

Subgen. I. CYLACTIS Rafin.

(Type R. saxatilis Linn.)

Shoots annual. Leaves 3-nate. Flowers bisexual. Fruit red to deep purple, drupels hardly coherent, parting from the receptacle. Receptacle flat or convex.

1. R. saxatilis Linn. 1753, Sp. Plant. 494; Syme, 1864, Eng. Bot. ed. 3, t. 441.

4x. A small, low plant. Barren stem forming offsets at the tip then dying back nearly to the base, from which rise the flowering shoots in the following year. Flower shoot erect, with a few, weak, straight prickles and simple hairs; terminal leaflet ovate-rhomboid-cuneate, incise biserrate; lateral leaflets subsessile; petiole furrowed. Flowers very few; sepals lanceolate, acuminate, patent; petals 5, glabrous, narrow, erect, white; stamens long, the interior filaments incurved at the apex. Fruit red; drupels 6 in the terminal, 2 in the lateral flower; seed slightly wrinkled or pitted, or nearly smooth. Flowers May-June. Calcicole. Woods and wet rocks.

E., W., S., I. Absent from the midlands and south England, and from central and south Ireland. All Europe except the south-west. Asia to the Altai Mts, Armenia, N. Persia and Kashmir.

2. R.arcticus Linn. 1753, Sp. Plant. 494; Smith, 1806, Eng. Bot. ed. 1,1.1585.

2x. A very small plant without a barren stem and without prickles. Leaves simple, 3-lobed, or 3-nate and rarely 5-nate; terminal leaflet ovate, cuneate, unequally serrate; lateral leaflets sessile. Flowers 1-3, showy, 1-25- 2-5 cm., one terminal and others axillary, long-stalked; sepals and petals 5-7(-io), sepals greenish, glabrous; petals purple or rose, broad ovate, clawed, often toothed; filaments purple, erect, incurved at apex, equalling styles. Carpels pilose. Fruit red to deep purple, drupels numerous, seed faintly wrinkled. Grows usually among Poly trkhum on gravel or grit covered with bog-earth. Comes up as soon as the snow melts in May and rushes into flower when still only 3-5 cm. high.