æquante stricte cylindrico extus puberulo labiis quam tubus longioribus anguste ovato-oblongis antico breviter tridentato postico integro, filamentis labiis paullo brevioribus antheris superpositis muticis, stylo minute bidentato.

Hab. Damaraland; T. G. Een.

Folia 1.5-2.0 cm. (raro 30 cm.) long., 0.6-1.5 cm. lat., utrinque cystolithis copiose instructa; petioli ± 0·4 cm. long. Pedunculi 0.3-2.0 cm. long. Bractea antica 10 cm. long., postica 0.6 cm., ambæ virides. Bracteolæ floris perfecti 0.7 cm. long. Calycis lobi 0.6-0.7 cm. long., 0·05-0·09 cm. lat. Corollæ tubus 0-8 cm. long., 0.2 cm. diam.; labia 1·15 cm. long. Filamenta 0·9 cm. long.; antheræ 0.1 cm. long. Ovarium 0.1 cm. long.; stylus puberulus, 1.5 cm. long. Capsulæ valvæ oblongæ, puberulæ, 0·65 cm. long. Semina 0.275 x 0.2 cm., levia, castanea.

Known from D. maculata Nees chiefly by the small leaves, the laxly arranged spikelets, and the shape of the small bracts.

[CORRECTION.-Pentanisia spicata (vide p. 38) is withdrawn, the plant described under that name being Otiophora scabra Zucc. Mr. Britten drew my attention to this mistake, the result of an error of observation with regard to the position of the ovule in the cells.]

SPARTINA TOWNSENDII.*

BY OTTO STAPF, Ph.D., F.L.S.

GIVING evidence some time ago before the Royal Commission on Land Erosion, Lord Montagu of Beaulieu called attention to the rapid spreading of a grass on the mudbanks of the Hampshire coast. According to him, it was accidentally introduced from the Argentine not many years ago and locally known as Rice Grass or Sea Rice. It was a rapid grower, overrunning mudbanks which had been hitherto bare and exposed, solidifying and raising them. The area covered by the grass was estimated at six thousand to eight thousand acres. The matter was submitted to the Director of the Royal Botanic Gardens at Kew, and subsequently I was invited to investigate the subject from the scientific point of view. Having during the last few months paid a number of visits to various points on the Hampshire coast and in the Isle of Wight to study the question in the field, I propose to give here a preliminary sketch of the history of the grass and the present extension of its area, adding at the same time a short technical paragraph for those who wish to make themselves familiar with the grass and the allied species which occur along with it.

[* Reprinted by permission from The Gardeners' Chronicle, Jan. 18th, 1908, where the paper is further illustrated by two maps showing distribution.] See the article on "Mudbinding Grasses "in Kew Bulletin, 1907, No. 5, pp. 190-197.

The grass which Lord Montagu had in view is Spartina Townsendii, a member of a genus numbering about eighteen species, mostly natives of America. With few exceptions, they inhabit sea marshes and muddy foreshores, under favourable conditions cover

ing hundreds and even thousands of acres. Four species are known to occur in Europe. Two, S. juncea and S. alterniflora, were introduced from the Atlantic coast of America, probably during the first half of the last century. S. juncea is confined to the western basin of the Mediterranean.

The other, S. alterniflora, was discovered by Loiseleur in the estuary of the River Adour near Bayonne in 1803, and then in 1829 by Borrer in the Itchen River near Southampton. A very complete account of it as it appeared there in 1836 was given by Bromfield. It has since then spread to some distance north of Northam Bridge in the Itchen River and to the Southampton Water as far as the Titchfield River on the eastern, and from Hythe to Redbridge on the western bank and from there to Millbrook. In France the grass has extended its area over a coast line of about twenty-five miles from Cap Breton (Landes) to the estuary of the Bidassoa River.

A

Of the three remaining species, one, S. stricta, has been known for a long time (since 1629), and is beyond doubt truly indigenous in Europe. It is found in England along the east coast from southern Lincolnshire to the Thames and on the south coast from Chichester to the Solent. On the Continent it occurs along the Atlantic coast from the estuary of the Schelde to near Gibraltar, and in a detached area at the head of the Adriatic. continuation of the Atlantic area is found on the coast of Morocco from Tangier to Mogador. The other two species are S. Neyrautii and S. Townsendii. S. Neyrautii was discovered by Neyraut near Hendaye in the estuary of the Bidassoa River about fifteen years ago, and was described by Foucaud (1894), who suggested that it was a hybrid between S. alterniflora and S. stricta, among which it is found growing.

S. Townsendii was first recorded by the brothers H. and J. Groves in 1879 from Hythe in the Southampton Water; but we have evidence that it existed there as early as 1870. According to the brothers Groves, it was already in the 'seventies rather common on both sides of Hythe Pier. This for years remained the only station. In 1883 it had not yet travelled beyond Cracknore Hard (two miles north of Hythe). In the Isle of Wight it was observed in 1893 (Yarmouth) and 1895 (Medina River); but nearly all other first records date from 1900 or after. To what extent it has spread during the last seven or eight years can be seen from the map [in Gard. Chron. l.c.]. The area thus conquered by the aggressive newcomer extends at present over a coast line the extreme points of which are over fifty miles distant. It would be tedious to trace the advance in detail; a few instances may suffice. In 1893 [Rev. E. F.] Linton found "several strong patches" of it near Yarmouth on the road leading from that place to Freshwater.† To-day it completely covers the mudbanks in the River Yar; it invades the adjoining marshland, and scattered clumps may be seen as far up as Freshwater Church. A few years ago, Lord Montagu assures me, there was no trace of it in the Beaulieu River; now it predominates everywhere to beyond Buckler's Hard, to quote from a manuscript report by Mr. J. F. Rayner, of

*Bromfield in Hooker's Companion to the Botanical Magazine, vol. 1, pp. 254-263, partly reprinted in Kew Bulletin, 1.c.

† [Report Bot. Exch. Club for 1893, 427.]

Southampton, "not only fringing the water, but running along every dyke, filling every pool and invading the broad borders of

SPARTINA TOWNSENDII (natural size).

marshland," and its advanced posts stand within half a mile of Beaulieu village. In 1895 and 1896, so Cosmo Melvill tells us,* there was none of the grass visible on the marshes and mudflats

between Hurst Castle, Milford, and Keyhaven; but in 1905 it was "plentiful and evidently rapidly increasing." On the roadstead of Poole Harbour a single small clump was discovered by ManselPleydell in 1899. Six years later Riddlesdell found it "in some quantity" by the fever hospital at Poole, whereas Mr. W. J. Goddard describes it in a letter to Col. Prain, dated October 8 of last year, as occurring in hundreds of big clumps all round the harbour on nearly every mudflat.

To explain the sudden appearance of the grass three theories suggest themselves. It may, like S. alterniflora, have been introduced, as Lord Montagu thinks; but so far no Spartina corresponding to S. Townsendii has been found in America; or it may have originated by way of mutation. It could only have sprung from S. stricta, which formerly occurred in Southampton Water; but S. stricta is little given to variation, and the differences are not of a character to support this theory. There is finally the hypothesis of the hybrid origin of S. Townsendii. This idea is not new; it was suggested by Foucaud in his note on S. Neyrautii. According to him, the latter was a hybrid of the formula S. alterniflora stricta, whilst S. Townsendii was S. stricta × alterniflora. Two circumstances lend considerable strength to that view; first, the fact that S. Neyrautii as well as S. Townsendii actually combine not a few of the distinctive morphological and anatomical characters of the supposed parent species, and, secondly, their occurrence just in the two-and the only two-parts of the world where S. alterniflora and S. stricta have met, namely, at the head of the Bay of Biscay and in the Southampton Water. This coincidence is very remarkable and has almost the demonstrative force of an experiment, the more so as S. Neyrautii and S. Townsendii approach each other so closely that they could not stand as distinct species if one wished to leave the theory of their hybrid origin out of consideration.

But another question, perhaps of more practical interest, presents itself. What are the conditions that enable the grass to spread with such amazing rapidity and get so firm a hold? The dispersal is no doubt mainly by seed. The grains fall with the spikelets, which float and would be carried about by the tides and currents until they are left on the beach or get caught somehow on the mudbanks. The grass does not seem to seed very freely, although it flowers profusely; but a few fertile clumps, as I have seen them, would after all give a good supply. When the seeds germinate, under natural conditions, we do not yet know. Possibly they behave like those of Zizania aquatica (a gregarious aquatic grass of North America), which lie in the water over the winter and germinate in the following spring. The seeds of these two grasses are remarkably similar, although the grasses are not allied at all. They have a highly developed embryo, which, in S. Townsendii, is considerably larger than the endosperm and bright green throughout, including even the leaf-like scutellum, which suggests that the process of germination passes off very quickly and effectively. Once established, the seedlings would soon grow