If sown when ripe, seeds of brambles germinate readily in open soil from soon after mid-March until June. This has been my invariable experience over more than 40 years, and it agrees fairly well with that of Babington who says that more than forty kinds sown at Cambridge in the autumn usually appeared in the succeeding May or June. Genevier, however, at Angers in the west of France says that in his experience the seeds are difficult of culture and take two years to germinate.

Seedlings are to be found in nature near blackberry bushes wherever birds perch, at the foot offences or hedges as well as on bare soil under the bushes themselves. They are recognized by the elliptical, ciliate seed-leaves. The seedlings will perish if they are too much exposed to the sunshine.

Species foreign to my neighbourhood which I have brought into my garden at Bickley I have subsequently met with near fences beside footpaths up to a distance of 200 yds. away, which gives an idea how far they may be transported by resident birds in autumn. The seeds are carried endozoically by birds that eat the berries, as the Turdidae, Magpie, Crow, Golden Oriole, mountain birds as Grouse, Blackcock, Ptarmigan, Capercaillie, and by mammals as Man, Bear, Fox and Pine Marten. They may be carried also epizoically overland by birds, especially by Woodcock, which, with wet mud on its legs after a night's feeding in marshy ground, repairs at dawn to crouch and doze under bramble bushes and hollies during the daytime. Woods called ‘ Woodcock Wood’ or ‘Cockshoot Wood’, or woods near them, especially on rather high ground, consequently often contain far-brought bramble species.

About 20 years ago a bush of R. glanduliger came up under a holly on Chislehurst Common at the highest point of the Tertiary plateau, where Woodcocks had been twice observed. Until then it was known only in west Sussex. As I used to pass the spot every week and had listed all the species, now 19, growing there on the green, I could be certain about its introduction.

A s would be expected, the fruits of different species of Rubus differ much from one another in edible quality: the best of them, allowed to ripen completely under the protection of a net or muslin, are well worth cultivation in the garden for home consumption. They do not travel well.

Mrs Lankester in Syme's English Botany praises the fruit of the dewberry as ‘very superior, and larger than any other blackberry’, an estimate which one cannot endorse, either as to flavour or size. She is possibly thinking of R. Balfourianus, of which one does sometimes meet with a strain producing large fruit of a mulberry flavour, and which Bagnall described as much more delicious even than R. gratus. It is perhaps the ‘Rubus morus, the Mulberry Bramble, so called by the Country People at Sutton, in Essex’. (Dillenius, Indiculus, appended to Ray's Synopsis (3rd ed.), extracted from Merret's Pinax (1666), p. 106.)

Babington, who cultivated blackberries in the Cambridge Botanic Garden, placed R. hastiformis amongst the best flavoured known to him. I have not grown it, but it was chosen at Merton for crossing with R. inermis Willd. and a useful fruit has been selected which is being extensively grown.

It is, however, within the power of anyone, without needing to know the name of any bramble, to choose a large fruit that pleases him from a bush on which the fruits are consistently large, sow the seeds at once, and in three years’ time begin to enjoy crops of the fruit, probably somewhat improved by cultivation, from his own garden. Not the smallest merit of the blackberry is that it improves the apple pie, pudding or stew, and is itself improved by the association.

Rubus is a remarkable genus in several respects. It has spread to all the continents; it shows signs of decay in its reproductive system whilst it is still in process of evolution, and so offers an opportunity of observing the origin of species. The best approach to a study of the genus is to try to identify the brambles met with, and in the endeavour, knowledge comes. The present work includes descriptions of all the forms found in Great Britain, Ireland and the Channel Isles, three-quarters of which are also found on the Continent.

On account of their relatively slow methods of distribution, their hardiness and the consequent preservation of old forms, the Rubi afford perhaps an unrivalled instrument for ascertaining the movement and succession of florulas and floras, if they are studied in connexion with the climatic and geological changes of the past.

If they may be comparatively unknown to most persons, they are nevertheless no strangers to any of us. From their wild habitats—woods and heaths, mountains and moors—they come to our hedges and garden fences, offering their flowers, like little roses, and then their fruit to our regard, after the cultivated bush fruits are all over. They readily accept the hospitality of a garden, and if taken in hand will be found, some of them, to yield delicious fruit if they are protected from the thrushes and allowed to ripen.

The principal object of the present work is to furnish the means of identifying any native species met with, and the means of understanding the causes and types of variation, and in some measure the probable natural relationships of the species. It has to be remembered, however, that the study of Rubi is still in an early stage of investigation, and there is room for much pioneering work in many directions.

A conspectus of species is prefixed to the descriptive part of the work, analysed to small ultimate groups of species, which are placed in order of frequency within their group. In addition the very common species are distinctively marked, and those other geographically restricted species are also suitably indicated. It is believed that these and other novel aids adopted in the main descriptions will speed identification and serve better than any other scheme.

Certain variable characters are found that are not known to be affected by any environmental condition and are presumably under genetic control. They are of a rather minor nature, and their presence or absence does not obscure the identity of the species. They are not due to hybridization; that is, they are not directly due to any fresh hybridization. Their appearance or disappearance in a new generation is apparently due to a recombination of genes that regularly takes place prior to the formation of gametes.

In some species one or other of these characters seems to have become fixed, and may be reckoned as one of the distinguishing characters of the species. In other species the same character may occur sometimes, not always. In yet other species it may not have been met with at all. For instance, pilose anthers occur always in Balfourianus, danicus, Drejeri, gratus, mucronifer and sciocharis; often in affinis, plicatus, regillus, silvaticus, vestitus and Winteri; never in adscitus, cardiophyllus, foliosus, pallidus, rudis and Sprengelii.

A survey is needed to learn in which species the following characters appear, and whether always.

1. 1a. Leaves 3-nate to 4-nate-pedate or 5-nate-pedate.

2. 1b. Leaves 3-nate to 5-nate-digitate or 5-nate-subdigitate, never 6-nate or 7-nate.

3. 1c. Leaves 3-nate to 5-nate-digitate and sometimes 6-nate or 7-nate.

4. 2. Number of sepals and petals where the number anywhere exceeds 5.

5. 3. Petals glabrous on the margin, or pilose.

6. 4. Anthers pilose; always?

7. 5. Carpels glabrous; pubescent; pilose; bearded.

8. 6a. Receptacle glabrous; or pilose.

9. 6b. Receptacle hirsute at the base with a brush of hairs protruding below the lowest carpels all round.

To observe the bush, the best way is to take a freshly opened flower and thumb down the nearer stamens outwards. This will give a clear view of the young carpels also. Receptacular hairs often grow out between the carpels and must not be mistaken for hairs on the carpels. To remove any doubt it is wise to take off a few of the upper carpels for examination; this will also expose the receptacular hairs clearly.

A good deal of information has been collected under the above heads, and some of it is included in the descriptions. Until it is much more complete it will be unsafe to begin to draw conclusions.

Goats, certainly, and (?) sheep also, when they are tethered up, browse brambles, and I have seen bushes that have been entirely stripped of leaves by them. On the Continent some buprestid larvae are very destructive, and attempts have been made in New Zealand to control their weed RUBI by liberating imported buprestids upon them.

Attacks of the rust fungus, Phragmidium spp., occur on the leaves and stem of brambles, but are not serious. More damaging in Canada is the fungus Hapalosphaeria deformans which destroys the anthers of cultivated blackberries such as R. laciniatus and the Boysenberry and spoils the fruit. For some thirty years past it has been found on R. latifolius in Scotland. The anthers become first granulated and then powdered with the white spores.

A virus is probably the cause of arrested, tufted, dwarf growth, in which the deformed panicles do not flower and the plant dies off in a few years. I can find no fungus on such bushes.

The most prevalent and most troublesome disease, however, found upon brambles is the felt-disease, which is believed to be a kind of gall due to a mite-attack. An extremely dense coating of brassy coloured felt extends over branches, leaves, pedicels, calyces and fruit; it masks or replaces the true felt or hair covering and inhibits the production of glands and acicles. The disease seriously checks the growth of the bramble and gives it a very sickly look. Some parts of the plant may be found to be less attacked than others, but it is better not to collect from such bushes at all. Rooted tips carry the infection, but if a ripe or nearly ripe fruit can be found and the seeds are cleaned thoroughly before sowing, clean seedlings can be obtained.

One does not know much about brambles, until one has mastered the range of changes a bramble plant undergoes due to age, season, climate, exposure, soil, habitat and so on; neither is it possible for a given bush or specimen to be named intelligently by comparison with another specimen or figure unless one knows how to allow for these modifying influences. One may instance particularly the deep shade of woods, an excess or deficiency of soil moisture, atmospheric moisture, a spell of hot weather, a run of cool rain, a shallow soil overlying rock or iron pan, exposure to salty sea breezes, a frost in May, lopping in hedgerows. The effects of these factors are discovered by observation of the brambles growing under such conditions. That they are of practical concern a few instances may prove.

R. carpinifolius sometimes grows in marshes and is then apt to make lax panicles and somewhat laciniated leaflets. Such a specimen was named R. leucandrus by Focke, and another R. carpinifolius var. laxus by Sudre. R. lacustris Rogers found in Lakeland is a wet-soil form of R. Lindebergii. R. vestitus growing in woods so far deceived Babington that he published it as a new species, R. Leightonianus, and was corrected by Leighton. Salter was only stopped by Borrer from describing wood-land R. vestitus as a new species, R. rotundifolius. R. ericetorum var. cuneatus Rogers & Ley is the same as R. ericetorum of Rogers. R. ericetorum ssp. sertiflorus var. scoticus Rogers & Ley is the same as sertiflorus of Rogers. R. botryeros Focke, at first associated with R. longithyrsiger by both Focke and Rogers, is the same as R. oegocladus of Rogers. Rogers repeats R. dumnoniensis Bab. under the name of R. cariensis Rogers (not Genev.). Sudre has this bramble under four names. The bramble which Rogers has as.R. Borreri, Sudre has figured under that name and again under R. retrodentatus; he has described it also as R. Schmidelyanus var. breviglandulosus.

These faults are due to remissness in observing the behaviour of the species in different environments.

Wherever known, the chromosome number is shown in the description. In some cases counts have been made only on one bush; the chromosome number is then starred. There is a possibility that a different number will be found when a second or third bush is examined in these cases, as where two bushes have been examined two different numbers have several times been obtained.

Certain results are, however, already sufficiently clear, namely (i) there is only one basic number, 7; (2) the chromosomes are found always in exact multiples of the basic number, with exceedingly rare exceptions; (3) the prevailing number is 28 in body cells, 14 in germ cells.

Three dwarf seedlings which appeared in a family raised from R. thyrsiger self-pollinated each had 27 chromosomes: it was claimed that they represented the systematists’ var. parvifolius. These dwarfs, however, proved to be comparatively infertile, whereas natural dwarfs are quite fertile. Moreover, several different natural dwarfs have now been studied and have been found to possess 28 chromosomes each. The implication, tfyen, based on the thyrsiger seedlings refers only to that case; the cause of the dwarfing in the other cases remains unexplained. It is also not known whether dwarfing is a reversible condition.

Different opinions prevail as to whether the forms of chromosomes can be distinguished well enough to throw light on the origin and relations of species in Rubus. In R. arcticus, idaeus, caesius, saxatilis, and hybrids artificially created between some of those species, the chromosomes have been studied, distinguished, delineated, and conclusions have been drawn from their form; their pairing behaviour in meiosis has also been established; but there is a difference of opinion between cytologists as to whether allotetraploidy or autotetraploidy is indicated in the parent species concerned, R. caesius and R. saxatilis. No doubt in course of time the doubt will be resolved to the satisfaction of cytologists. From a morphological point of view it seems very probable that R. caesius and R. saxatilis are of hybrid origin, and that one parent in each case is R. idaeus