Involucri squamæ tenuiores laxiores, fere innocua: caules 3-8

pedales capitula nuda.

C. EDULIS. Laxe tenuiter araneosus, mox glabrescens viridis ; caule folioso; foliis sæpius sinuato-pinnatifidis membranaceis; capitulis subpaniculatis vel subglomeratis; involucro arachnoideo; corollis purpureis (quandoque pallidis?) gracilibus subinæqualiter 5-fidis, lobis demum filiformibus incrassato-apiculatis. Cirsium edule Nutt. 1. c.From British Columbia southward along the coast of California to the Bay of San Francisco. Filaments sometimes hairy, as described by Nuttall, very commonly glabrous: perhaps a subsexual difference. Dwarf specimens collected by Lyall in the northern Cascade Mountains have the leaves pinnately parted into narrow divisions.

C. REMOTIFOLIUS. Caule sæpius parce foliato; capitulis subpaniculatis; foliis subtus araneoso-dealbatis raro denudatis pinnatipartitis, lobis angustis; involucro tenuiter arachnoideo glabrescente, squamis lineari-attenuatis; corollis ochroleucis inæqualiter 5-fidis, lobis 3 vel 4 altius coalitis. Carduus remotifolius Hook. 1. c. Cirsium remotifolium DC. C. stenolepidum Nutt. 1. c. - Oregon near the coast to Humboldt Co., California.

§ 3. Involucrum subglobosum gradatim imbricatum, squamis plerisque scarioso- vel fimbriato-appendiculatis: flores albidi vel flaviduli. (Echenais Cass., DC.)

C. PARRYI. Viridis, vix araneosus; capitulis subracemosis parum nutantibus; foliis lanceolatis sinuato-dentatis; involucro laxius imbricato; squamis subchartaceis, exterioribus paullo brevioribus linearibus seu lanceolatis, marginibus superne tenuiter scariosis pectinato-fimbriatis ciliatisque, intimis appendice scariosa lacera parva superatis ; corollis flavidis, lobis fauce longioribus. - Rocky Mountains of Colorado Territory, at the elevation of 8-9,000 feet, coll. Parry (no. 34), and Hall and Harbour (no. 340), which I had doubtfully referred to Cirsium edule in the account of their collections: also Vasey, no. 350, referred to Echenais carlinoides, and Wolf and Rothrock, no. 460. Hall and Harbour's no. 341 is probably a hybrid of this with C. eriocephalus. In foliage this resembles C. (Echenais) Sieversii. In the involucre, &c., this species and some forms of the next offer a complete transition between Echenais and the foregoing section, and they seem to be further connected by hybrids or by variations. The delicate fringe of the outer involucral scales is usually pectinately dissected into almost setiform divisions, often passing into or mixed with long and soft jointed

hairs; the terminal prickle weak. Subtending bracts spiny-ciliate, the inner passing into the involucral scales.

C. CARLINOIDES Schrank, var. AMERICANUS.

Rocky Mountains of Colorado Territory, Hall and Harbour (no. 342), E. L. Greene; Western part of California, Samuels, Bolander: forms with short and broad scarious and lacerate appendages to most of the scales of the involucre, tipped with an extremely short prickle, and few or no prickly-fringed subtending bracts. Also, Mendocino Co., California, Kellogg, a form with exterior involucral scales hardly at all appendaged, and the inner with rather small acuminate appendage,― possibly a hybrid with C. remotifolius. This, or a form like it, appears to be Cirsium scariosum Nutt. in Trans. Amer. Phil. Soc. 1. c. p. 420, from the plains of the Rocky Mountains, which was accidentally omitted from the Flora of North America.

II. Notes on Borraginaceæ.

I am

COLDENIA Linn. Upon a revision of the plants of this group, the more convinced that the genus Coldenia should have the extension which I proposed in Proc. Am. Acad. 5, p. 340, and should include Ptilocalyx Torr. also. And it is pleasant to note that a genus which was dedicated to one of our worthies of the colonial period, has proved to be mainly American, although founded on an Indian plant. The section which I proposed, under the name of Tiquiliopsis, if strengthened on the one hand by a second species (C. Palmeri Gray, Proc. Am. Acad. 8, p. 136) as respects the corolline appendages, is invalidated on the other by the discovery that its embryo does not accord with that of T. Nuttallii; but it is still unlike that of Tiquilia. Mr. Watson, in redescribing the T. Palmeri (in Bot. King, p. 247), states that the tube of the corolla is "without scales at the base." There are not, indeed, such free scales as those of T. Nuttallii, but in their stead are much longer and salient plicæ, reaching up to the insertion of the slender filaments. The fruit, which Mr. Watson first made known, he describes as of "a single obovate-globose smooth nutlet, attached at the base, and without ventral sulcus." There are often two such nutlets matured; but the rounded scar is ventral, not basal, yet very different from that of T. Nuttallii. Of albumen there is barely a trace. The character "cotyledons rounded, flat, entire, incumbent upon the shorter radicle," is correct, except that they are rather hemispherical than flat. This turning up of the rather long radicle upon the back

of one of the thick cotyledons is most peculiar and remarkable, and is strikingly in contrast with the deeply hippocrepiform cotyledons of T. Nuttallii, surrounding the radicle, as represented in Dr. Torrey's plate. I am disposed to keep up the section Tiquiliopsis for these two species with anomalous embryos and appendaged corolla-tube.

HELIOTROPIEÆ Fresenius. The proper stigma in Heliotropium and Tournefortia, occupying the margin of an annular or peltate disk, and surmounted by an appendage which has generally been taken for stigma, recalls the similar structure in Apocynaceæ.

HELIOTROPIUM Tourn.* If three genera are to be admitted in the

*The North American species as now known are:

§ 1. EUPLOCA. Vide supra.

H. CONVOLVULACEUM Gray, 1. c.

§ 2. EUHELIOTROPIUM. Heliotropium & Schleidenia (Endl.) Fresenius.
*Orthostachys R. Br. (Preslæa Mart.). Schleidenia Endl., Fresenius.
Appendix stigmatis elongata, subulata. Folia angusto-linearia.

H. GREGGII Torr. Bot. Mex. Bound. p. 137. brevissimus.

Corolla limbo amplo: stylus

H. ANGUSTIFOLIUM Torr. 1. c. Corolla lobis ovato-lanceolatis: stylus gracilis. H. TENELLUM Torr. 1. c. Calyx inæqualis: stylus perbrevis.

Appendix stigmatis brevis conica.

H. LIMBATUM Benth., var. CONFERTIFOLIUM Torr. Bot. Mex. Bound. Hartweg's plant, on which H. limbatum was founded, has the aspect rather of H. hispidum than of Torrey's plant of our Mexican border, collected by Berlandier, Gregg, Wright, &c. But Coulter's no. 1051, and corresponding specimens by Dr. Edwards, &c., are intermediate. This var. confertifolium Torr. is very like H. microphyllum Swartz., as represented by Wright's Cuban no. 3139. This, however, has a much smaller corolla and a shorter style; and H. imbricatum Griseb. would seem to be a form of it.

H. glomeratum A. DC? H. bur-
Florida, Buckley, &c.

H. POLYPHYLLUM Lehm. Asper. & Ic. t. 8. siferum C. Wright in Griseb. Cat. Cub. p. 211. Var. LEAVENWORTHII (H. Leavenworthii Torr. herb.) is a strict and slenderleaved form; but specimens recently collected by Dr. Edward Palmer, in the same part of Florida where Dr. Leavenworth collected it, seem to have goldenyellow flowers!

H. PHYLLOSTACHYUM Torr. 1. c. 1859. H. myosotoides Chapm. Fl. p. 330, 1860. Berlandier's numbers 1538, 3038, referred to this species by Dr. Torrey, are more probably depauperate states of H. hispidum HBK.

** Flores ebracteati in spicis scorpioideis sæpius conjugatis vel 1-2-furcatis: antheræ liberæ.

H. EUROPÆUM Linn. Stigma appendice tenuiter subulata superatum. Naturalized southward.

VOL. X. (N. S. II.)

manner and upon the characters proposed by Fresenius in the Flora Brasiliensis, Nuttall's name of Euploca would by right of priority take the place of Schleidenia. But, in retaining the comprehensive Heliotropium of Linnæus (Piptoclaina of Don perhaps separated), I prefer to include Schleidenia, and also Bentham's section Platygyne, in Euheliotropium, and to make of Euploca a primary section. It is characterized by the long filiform style, strongly penicillate appendage to the stigma, and didymous fruit, which separates promptly into four almost hemispherical half-carpels. The corolla is unusually large for the genus. Many years ago we had the plant in cultivation, and thought very ornamental. It is desirable, and it ought not to be very difficult, to obtain it again. The pure white blossoms are open not merely at sunset, but also (according to my memorandum) for nearly the whole day. The name of H. convolvulaceum was applied to it in the Mem. Am. Acad. 6, p. 403, 1859. There is an equally slender style in Grisebach's H. serpylloides of the West Indies, and the cone surmounting the stigma is occasionally bearded in other species.

it

H. parciflorum of Griseb. Cat. Pl. Cub., and mentioned in the Flora of British West Indies, cannot be the Preslæa parciflora of Martius (Schleidenia Fresen.), for it has a strongly beaked fruit. It is perhaps merely a form of H. humistratum Cham.

H. fruticosum Linn., described by Grisebach as having "stigma as long as the style," has (even in a specimen named by and received from him) a style several times longer than the stigma and its tip.

Heliophytum molle Torr. Bot. Mex. Bound. p. 138, with globose, not at all didymous, and when fresh probably more or less drupaceous fruit, I refer to Tournefortia. =T. mollis.

LITHOSPERMUM Tourn. Characters from the insertion of the stamens and length of the style should be suspected; for the tendency to

H. INUNDATUM Swartz. Stigma cono obtuso capitatum.

H. CURASSAVICUM Linn. Stigma umbraculiforme, cono obsoleto.

§ 3. TIARIDIUM (Lehm. Heliophytum DC.)

* Fructus didymus, nuculis parallelis.

H. PARVIFLORUM Linn. Keys of Florida and along the southern borders of Texas.

H. GLABRIUSCULUM. Heliophytum glabriusculum Torr. 1. c. South-western borders of Texas.

Fructus mitræformis. - Tiaridium Lehm.

H. INDICUM Linn. Naturalized in Southern Atlantic States.

dimorphism, which is obvious in the Batschia section (although not demonstrated in all of the species), may pervade the genus."

MERTENSIA Roth. A new study of this genus enables me to make a few corrections to my paper on the American species, pub

* Our North American Lithosperma of the Batschia section may be character ized thus.

1. Corolla ut videtur pallide flava: folia floralia minora, calyces fructiferos haud superantia.

L. MULTIFLORUM Torr. in herb.; S. Wats. Bot. King (adnot.), p. 238. L. pilosum Gray in Sill. Jour. 34, p. 256, & in Proc. Acad. Philad. 1863, non Nutt. Colorado in the lower mountains to New Mexico and Texas. The throat of the corolla is nearly naked. Nuttall's L. pilosum is the same as L. ruderale Dougl. & Hooker. I was misled into taking this species for it through Nuttall's reference of his to Batschia, and his idea that the corolla was yellow, whereas in that species it is only yellowish in the manner of L. officinale.

2. Corolla aurea vel aurantiaca: folia floralia pleraque conformia, calyces superantia,

+ Oblonga vel ovata: corolla nec tubo prælongo, nec plicis faucialibus fornicatoinflexis.

L. CALIFORNICUM. Pube molli hirsutum; foliis lanceolatis; corolla flava, tubo calyce sesquilongiore, fauce ampliata 5-loba, lobis brevibus, plicis fauciali bus obsoletis, annulo ad basim tubi nudo. - L. canescens var. Torr. Bot. Whipp. p. (68) 124. Grass Valley, California, Dr. Bigelow. The anthers are high and the style short; but this is probably only one form.

L. CANESCENS Lehm. Pube molli primum canescens: corolla læte aurea, plicis faucialibus pubescentibus prominulis, annulo basilari nudo. - L. sericeum Lehm. must belong to this, but the synonym Anchusa Virginica Linn. is to be excluded. I do not know what the glabrous plant in the Linnæan herbarium is; but the Gronovian plant is an Onosmodium, Morison's is probably Lithospermum hirtum, which is the Puccoon of the Southern States, and Plukenet's may be either species.

L. HIRTUM Lehm. Pube hispida demum asperum; floribus majoribus; corolla læte aurantiaca, plicis faucialibus prominulis, annulo basilari dentibus 10 hirsutissimis instructo.-L. Bejariense DC. is of this species.

+44 Folia omnia linearia angusta: corollæ bene evolutæ tubo prælongo, plicis faucialibus fere fornicatis, lobis inciso- vel undulato-crenulatis.

L. ANGUSTIFOLIUM Michx. L. linearifolium Goldie? Pentalophus longiflorus & P. Mandanensis A. DC., cum syn. cit. etc. L. breviflorum Engelm. & Gray, Pl. Lindh. 1, p. 44. It is to Mr. Bebb (see Amer. Naturalist, 7, p. 691) that we owe the demonstration that the long-flowered species (L. incisum Lehm. &c.) is the perfect form, as we may say, which, later in the season and especially upon lateral shoots, goes on to produce depauperate flowers, with corolla and style hardly equalling the calyx, and without doubt of cleistogenous fertilization. In this state it is L. angustifolium Michx. - the earliest and an appropriate name.