ART III.-Chionophila Benth. A Morphological Study: by THEO. HOLM. (With 15 figures, drawn from nature by the author.)

The genus Chionophila Benth. is a member of the tribe Cheloneae, and its nearest allies in North America are Chelone and Pentstemon; as distinguishing characters Gray, Bentham and Hooker consider the structure of the calyx, corolla and fruit as the most important. The calyx being funnelform, obtusely five-lobed in Chionophila, deeply five-parted in the two other genera; the corolla being densely bearded at the base of the lower lip in Chionophila, while in Pentstemon this structure recurs only in certain species: P. barbatus Nutt., P. laevigatus Soland., P. pubescens Soland., and a few others; moreover the capsule being septicidally dehiscent in Chelone and Pentstemon, but at first loculicidal in Chionophila; in the last of these the capsule is enclosed in the marcescent calyx and corolla, while it is free in the others; also the seeds are different, being rather large, oblong, with a very loose, reticulated outer coat in Chionophila, winged in Chelone, and angulated, but marginless in Pentstemon. Very little, however, is said about the habit of Chionophila : Herba perennis, humilis, caespitosa, glabra is the characterization offered by Bentham and Hooker; a dwarf perennial, glabrous or nearly so is all, that Gray has to say. At the time of Bentham and Gray the genus was supposed to be monotypic, but since then a second species has been proposed by Professor L. F. Henderson1 C. Tweedyi, by Canby and Rose referred to Pentstemon, P. Tweedyi. However this second species differs from C. Jamesii Benth. according to Professor Henderson, by the corolla being quite saccate at base dorsally, thus causing the corolla to be turned aside at quite a strong angle with the axis of the inflorescence, beside by the calyx being short and more lobed, and the lower lip of the corolla being merely papillose. Professor Henderson found this new species in open, loose soil at the bases of mountains in Idaho. C. Jamesii, on the other hand, is a member of the high alpine flora of the Rocky Mountains of Colorado and Wyoming.

'Henderson, L. F. New Plants from Idaho and from other localities of the North West. (Bull. Torr. Bot. Club, vol. 27, 1900, p. 352.) 'Canby and Rose in Bot. Gazette, vol. 1o, 1896, p. 67.

Recently C. Tweedyi has been raised to generic rank, receiving the unfortunate name "Pentstemoniopsis," for as will be shown in the subsequent pages, the plant is undoubtedly just as close an ally of Chionophila as of Pentstemon; moreover the frequent use of making generic names by adding "opsis" has resulted in names of too great length, and sometimes also in such terrific combinations as: Saxifragopsis, Stellariopsis, Bombacopsis etc.

Now with respect to Chionophila Jamesii the plant is not caespitose, but stoloniferous, and the shoot above ground shows very plainly the structure of a typical monopodium. There is a rosette of leaves, and generally only one flower-bearing stem, which is axillary. In the center of the rosette is a bud, vegetative, which terminates the shoot, and the flower-bearing stem is always axillary, proceeding from near the base of the shoot, from the axil of one of the basal leaves. The leaves are opposite with long sheaths, and the accompanying diagram (fig. 1) shows the structure, which is the commonest met with. Four pairs of leaves constitute the rosette, L.-L.; of these L. enclose a vegetative bud, which will develop a new rosette in the coming year, and in the axil of one of the leaves L.' is a floral stem (St.) When stolons develop, they generally proceed from the axils of the leaves of the previous year, now present as withered leaf-sheaths, and these stolons consist of two to three stretched internodes, covered by leaf-sheaths; they grow in a horizontal direction, and become terminated by a rosette of leaves, repeating the structure of the mother-shoot. During the first season the stolons remain attached to the mother-shoot, and roots develop freely from the nodes. The number of stolons may vary from one to as many as six, but one or two are the most frequent. The subterranean stem is ascending, densely covered with remnants of leafsheaths, and of a soft, fleshy consistence.-The monopodium represents a structure, which is evidently rare; we remember this structure being characteristic of certain species of Carex, for instance C. laxiflora, C. digitata, C. Fraseri and some others, which I have described in this journal. A sympodium, however, is the most frequent structure observed in rhizomes especially. Among the Smilaceae Smilax herbacea shows the interesting case of Rydberg, P. A. Flora of the Rocky Mountains, New York 1917.

the rhizome representing a sympodium, while the aerial stem is a true monopodium. Unfortunately the literature gives little information about these structures, and with regard to the structure of the shoot above ground being a monopodium so distinctly represented by Chionophila, there is no mention of this structure being represented by any of the other Scrophulariaceae; as a matter of fact the word "monopodium" does not occur in any of the American manuals of Botany, dealing with floras of certain regions. Nevertheless it constitutes a character of no small importance, and none of the species of Pentstemon, which I have examined, exhibit this structure. It may be mentioned at the same time, that Arabis dentata exhibits a monopodium exactly as that of Chionophila, and is the only Arabis, known so far, of which the shoot is of monopodial growth. The monopodium occurs also in Claytonia, Calandrinia, Viola etc., which I have described in previously published papers on these genera.

Concerning the inflorescence in Chionophila, this is not a spike, but a unilateral cyme, a monochasium; the floral stem begins with two or three pairs of small, linear, opposite bracts without buds; but above these the bracts become somewhat larger, broader, and we find at each pair a single flower, and a lateral branch, which again terminates in a flower with two bracts and a lateral branch and so on. In Chionophila the composition of the inflorescence is not so very distinct on account of the short internodes, the leaves and flowers being rather crowded, but in Collinsia grandiflora Dougl. the structure is plainly visible, and corresponds exactly with that of Chionophila. In Pentstemon, on the other hand, there is a regular cyme in the axil of each leaf; in P. Newberryi Gr. the axillary cymes are one-flowered, there being only a single flower, but with two bracts, indicating a "dichasium" with two lateral branches suppressed, non-developed.

Characteristic of Chionophila, when compared with Pentstemon, is thus the monopodial ramification of the shoot, and the inflorescence representing a monochasium. The floral structure has been described very exactly by Gray, Bentham and Hooker, except that it seems to have been overlooked, that the calyx, in fresh specimens, is distinctly folded lengthwise as shown in figure 2.

Internal structure of the vegetative organs.

The roots:-In Solereder's treatment of the ScrophuAM. JOUR. SCI.-FIFTH SERIES, VOL. I, No. 1.-JANUARY, 1921.

lariaceae nothing is said about the root-structure. With respect to C. Jamesü the structure is of special interest, since I observed the development of cork to take place in the peripheral stratum of the cortex, the stratum bordering on the exodermis (fig. 5), and, as we remember, De Bary records only two cases of cork being developed in the peripheral strata of the root-cortex, namely in Clusia and Bignonia capreolata ;* since then Olivier detected a similar structure in Artanthe pothifolia, Jasminum humile, and Ruyschia Souroubea; to these may be added Tecoma radicans and Cephalanthus occidentalis."

There was no primary root in any of the specimens, which I collected, and the root-system consisted merely of several strong, secondary roots developed from the subterranean part of the stem; these roots were sparingly branched, and only slightly hairy. Inside the epidermis is a thinwalled, contractile exodermis (Ex. in fig. 5), covering an open cortical parenchyma of about twenty layers, with no starch or crystals; a phellogen (Ph.) arises, as stated above, in the peripheral stratum. Inside the cortex follows a thinwalled endodermis (End. in figure 6), surrounding the pericambium (P), in which tangential cell-divisions appear outside the protohadrome, being the only indication of secondary increase in the young roots. The roots are pentarch with five strands of leptome alternating with five short rays of hadrome. The center of the stele represents a relatively broad, thinwalled pith without starch.

As the roots grow older the cortex collapses, and some few secondary vessels develop on the inner flank of the leptome. The main stem, which is terminated by the rosette of leaves, shows the following structure. The upper internodes are cylindric, glabrous and smooth. A thin cuticle covers the slightly papillose epidermis, and the cortical parenchyma is homogeneous, of about twenty strata with distinct intercellular spaces, and containing a little starch. There is no endodermis, and no pericycle, but a band of isolated, small strands of leptome, and a continuous band of cambium, inter- as well as intra-fascicu

De Bary, A. Vergleichende Anatomie, p. 563. 1877. See also Van Tieghem, Ph. Recherches sur la symétrie de structure des plantes vasculaires. (Ann. sc. nat. Bot. Ser. V, 13. p. 258. 1870.)

Olivier. Recherches sur l'appareil tégumentaire des racines, ibid. Ser VI, 11. p. 124. 1880.

Holm, Theo. Rubiaceae; Bot. Gaz., (vol. 43, p. 155). 1907.

lar; there are eight primary mestome strands with a few vessels, but with neither mestome-parenchyma nor libriform; the center of the stele contains a thinwalled, starchbearing pith. Further down, and below the rosette, the epidermis is more or less collapsed, and replaced by two to three layers of cork, developed from the hypodermal structure of the cortex. The cortex is as above; in the stele is no endodermis, but a parenchymatic pericycle, and two concentric bands of leptome, separated by some ten layers of radially arranged parenchyma, a secondary cortex. Some secondary vessels may be seen in the primary mestome-strands, and the hadrome thus constitutes much deeper rays than in the internodes above.

The flower-bearing stem, which is axillary, is twowinged at the base, with the stele circular (in cross section). It is glabrous and smooth, but the cuticle is, here and there, wrinkled longitudinally. The epidermis is thickwalled, especially the outer cell-wall, and the cortex is also somewhat thickened, and consists of six to seven strata, very open from the wide intercellular spaces, and destitute of starch and crystals. Endodermis is barely distinct (End. in fig. 7), and the pericycle is merely present as a single layer of small-celled parenchyma (P. in fig. 7), which is continuous and bordering directly on the numerous small strands of leptome. There are eight primary mestome-strands, containing a few (four) layers of radially arranged, thickwalled, and porous mestomeparenchyma. The interfascicular tissue consists of leptome, and three to four layers of radially arranged, thickwalled parenchyma. The pith is solid, homegeneous, with no starch or crystals (fig. 7). In the wings of the stem is only thinwalled cortex, no collenchyma being developed.

This portion of the flower-bearing stem was subterranean, but the structure of the upper, aerial internodes is about the same. However, small hairs with cuticular pearls (fig. 9) occur on the internodes, and the cuticle is more wrinkled. The epidermis is more thickwalled, the cortex contains chlorophyll, and is very open from lacunae of quite considerable width, but otherwise the inner tissues are as described above.

Characteristic of the stem is thus the complete absence of collenchyma and stereome, the only strengthening tissue being represented by the thickwalled, interfascicular parenchyma.