Astragalus canadensis L. var. canadensis

Authors

Rupert C. Barneby

Authority

Barneby, Rupert C. 1964. Atlas of North American Astragalus. Mem. New York Bot. Gard. 13(2): 597-1188.

Family

Fabaceae

Scientific Name

Astragalus canadensis L. var. canadensis

Type

“Habitat in Virginia, Canada."—Holotypus, labeled “Astragalus 3, Hort. Cliffort.,” LINN!

Synonyms

Astragalus carolinianus L., Astragalus canadensis var. carolinianus (L.) M.E.Jones, Astragalus canadensis f. monticola Gand., Astragalus oreophilus Rydb., Astragalus halei Rydb., Astragalus canadensis var. longilobus Fassett

Description

Variety Description - Variable in stature, but usually tall and robust, the stems either slender and almost solid or stout and hollow, (3.5) 4-12 (16) dm. long, usually branched or spurred at 1-several nodes preceding the first peduncle, sometimes simple, the herbage green or subcinereous, the leaflets commonly glabrous or glabrescent above, the longest hairs of the herbage up to (0.25) 0.3-0.7 mm., and of the inflorescence up to (0.4) 0.5-1.1 mm. long; stipules 3-18 mm. long, the lowest ones usually ruptured and irregularly circumscissile in age; leaves (3.5) 5-25 (35) cm. long, with (9) 15-35 broadly lanceolate, oblong, ovate, or elliptic, obtuse or mucronulate, or truncate-emarginate leaflets 0.6-4.5 cm. long, all paler beneath and carinate by the prominent suture; peduncles (2.5) 4-11 (13) cm. long, commonly ± half as long as the leaf; racemes usually dense, rarely looser and racemose, mostly 4-16 cm. long and (2) 2.5-3 cm. in diameter at anthesis; pedicels in fruit 1.2-2.5 mm. long; calyx (4.6) 5.5-9.6 (10.3) mm. long, mostly white-, rarely in part fuscous- or black-pilosulous, the tube sometimes glabrate, the tube (3.4) 4.1-6.7 mm. long, (2.5) 2.7-4.1 mm. in diameter, the teeth subulate orlinear-subulate, 1.2—4 (5.1) mm. long; petals greenish-white or dull straw-colored; banner 11.3-16.3 mm. long, 5-8 mm. wide; wings 10.1-14.5 mm. long, the claws (4.2) 5-7.1 mm., the blades 7-9.6 mm. long, 1.2-2.2 (2.5) mm. wide; keel 9.3—13 mm. long, the claws (4.6) 5—6.8 mm., the blades 4.9—7 mm. long, (2.2) 2.5—3.4 mm. wide; body of the pod (9) 10-15 mm. long, 4—5.2 mm. in diameter, terete or nearly so but carinate ventrally by the suture, the nearly always porrect cusp mostly 2-4 mm. long, the valves glabrous or more rarely puberulent, exceptionally strigulose; ovules (18) 20-26 (27).

Distribution and Ecology - Low moist prairies, creek banks, roadside ditches, commonly in rich soils moist in spring, but also in dry rocky woodlands, on sandstone bluffs or cliffs, and on stony or sandy beaches and shores of lakes and rivers, westward on short-grass prairies, in mountain woodlands, and on stream banks, to the south sometimes in low sandy pine woods, mostly 10-3000 feet but ascending to 3500 feet in the Appalachians and to 7000 feet in Colorado, widely dispersed from southwestern Mackenzie and northern British Columbia to James Bay, Ontario, the St. Lawrence River, Quebec, and Lake Champlain, Vermont, south over the plains, through the Mississippi Valley (where commonest), and the Midwest to southwestern Texas, central Louisiana, northeastern Alabama, and along the Appalachians from West Virginia to northern Georgia, rarely east to northern New Jersey and the Coastal Plain in southeastern Virginia and the Carolinas; also along the east slope of the Rocky Mountains from Montana to northern New Mexico, and sparingly east of the Continental Divide in northern Idaho (Pend Oreille River, possibly introduced), Utah (Sevier Valley; Provo Canyon; Cache County; apparently native), and about the headwaters of the San Juan River in southwestern Colorado.—Map. No. 76.—May to September.

Discussion

The circumscription of var. canadensis has been drawn rather broadly in order to include all forms of the species with pod rounded and not sulcate dorsally, whether glabrous or puberulent. The entity so defined is polymorphic, individual plants and populations showing considerable variation in stature, number of leaflets, length and density of the racemes, proportions of the calyx, and length of the banner. There is evidently some correlation between certain types of variation and dispersal patterns, but no reliable method of separating named geographic races has been found. The phase of the Canada milk-vetch that is presumed to have been cultivated in France in the XVII century is the comparatively slender one found along the lower St. Lawrence River in Quebec, where it occurs on river banks and on gravel bars of streams. Here the stems are ordinarily about 4-7 dm. long, the leaflets are 7-11 pairs glabrescent, and relatively small, and the pod is consistently glabrous. The same phase or one identical so far as can be determined from herbarium specimens, has been traced westward in scattered stations across southern Canada and adjoining United States, principally in shore habitats, and reappears in mountain woodlands and on shady stream banks in the Rocky Mountains of Alberta, Colorado, and northern New Mexico. The plant of the southern Cordillera has been segregated as A. oreophilus, but the supposedly broad and short bracts are not a reliable differential character. The commonest phase of the Canada milk-vetch extends over a wide area from upper New York to North Dakota and south through the Mississippi Valley to the Gulf Coastal Plain in eastern Texas. It differs from its northern-montane counterpart in its extremely robust habit of growth (the stems mostly 5-12 or even 16 dm. tall), in having mostly 11 to 17 pairs of ample leaflets on the average more densely pubescent, and in the pod which varies from glabrous to puberulent. However, numberless examples illustrate a passage between the ideally conceived variants, and even the extreme types are geographically scattered (e.g., Marie-Victorin 3210, NY, from Jacques-Cartier County, Quebec, with up to 31ample leaflets, and many slender forms from as far south as Oklahoma). The specimens picked out by Rydberg as the basis of A. Halei are of the stouter type. But the spreading or reflexed pod, supposedly the special diagnostic feature of A. Halei, was described from a single specimen bearing fruits of the preceding season, battered by winter weather and empty of seeds, no form of A. canadensis with pods naturally reflexed is known to exist. Independently of variation in stature and amplitude of herbage, the calyx of var. canadensis varies a good deal in the proportion of teeth to tube. This is not simply a variation in length of the teeth alone, it is rather that the depth of the intervening sinuses determines the ratio of tube to tooth. A form with long teeth, especially frequent in the Great Lakes region, has been described as var. longilobus. In support of this proposal Fassett (Leg. Pl. Wis. 72. 1939) introduced a highly speculative history of the fortunes of A. canadensis in eastern North America during the close of the last Tertiary glaciation. According to Fassett’s hypothesis, a primitive stock characterized by long calyx-teeth and sparse pubescence must have survived as relic populations in the Appalachians (Jones’s var. carolinianus) and in Wisconsin’s Driftless Area; whereas a contemporary stock distinguished by shorter calyx-teeth and more copious pubescence (corresponding with today’s var. canadensis in a restricted sense) was obliged to retreat southward. As the ice sheets moved northward, the southern population followed and eventually meeting the Wisconsin relics merged with them to produce a distinct entity, var. longilobus, supposed to combine the pubescence of the immigrants with the long calyx-teeth of the old native stock.

Fassett based his hypothesis on assumptions which cannot be proved or are actually false. In the first place, the calyx-teeth of the so-called relic var. carolinianus are consistently long only in the literature, not in any natural range; I have recorded a variation of 1.9—3.8 mm. in plants of montane habitats ranging from West Virginia to Georgia. The teeth of var. longilobus, according to Fassett himself, vary from 2-3.7 (actually up to 5) mm.; but there are examples of teeth up to 3 mm. long from as far south as central Oklahoma, where the influence of the Wisconsin relics could hardly be expected to reach. Furthermore, if var. longilobus is the progeny of introgression between a carolinianus-like relic and a recent immigrant "var. canadensis," we must assume, in the absence of specimens from the Driftless Area and adjacent regions which combine the teeth of the latter variety with the pubescence of the former, that the tooth-character was consistently dominant and the pubescence-character consistently recessive. This point is no longer demonstrable now that the hypothetical Wisconsin relic has been replaced by hybrid offspring. It seems so much simpler (and as probably correct) to interpret the long calyx-teeth, which occur in scattered stations through a great part of the range of our inclusive var. canadensis, as no more than a simple variation coordinate with those found in many widely dispersed astragali in whose history glaciation can have played no part. Species of wide distribution generally exhibit statistically different modes and frequencies of variation in different parts of their ranges, and the fact that in A. canadensis relatively long calyx-teeth are prevalent in (although by no means confined to) the general region of the Great Lakes is neither especially surprising nor taxonomically significant.

Mention has already been made of a var. carolinianus, supposedly the A. carolinianus of Linnaeus (but see further, below), said to represent the Canada milk-vetch in the southern Appalachians. Like so-called A. oreophilus in the Rocky Mountains, the eastern mountain race is found along streams and in moist, deciduous woodland; it is commonly distinguished from var. canadensis of the Mississippi Valley and northward by a narrower and more open flowering and fruiting raceme, and the flowers at the same time are relatively small. There is something to be said in favor of recognizing an eastern montane variety, so long as we confine its distinguishing character to a loose raceme. It has been suggested (e.g., by Fernald, 1950, p. 911) that the calyx-teeth of var. carolinianus are longer than those of var. canadensis, but this is not the case; both calyx and flower can be matched, separately and together, in many individuals from lowland habitats in the middle Prairie States. It is worth recalling that Jones (1923, p. 168) originally defined var. carolinianus as ranging widely over the Southeastern States north at least to the Ohio River, and the concept of a montane Appalachian variety is comparatively modern.

The history of the Canada milk-vetch, the first astragalus described from North America, goes back to the latter half of the XVII century, when it was already under cultivation in Paris (Dodart, Mem. Acad. Roy. Sci. Paris 4: 257, cum tab. 1676). Old garden specimens are preserved in the Paris herbarium (P), but unfortunately none is dated. A form of A. canadensis, perhaps an independent introduction from the British colonies, was known to Dillenius in England about 1732; and Linnaeus saw it in Clifford’s garden at Hartekamp shortly after. There are two specimens of A. canadensis in the Linnaean herbarium. Gray (in T. & G., Fl. N. Amer. 1 (Suppl.): 649. 1840) recognized the sheet labeled "Astragalus 3, Hort. Cliffort." as genuine A. canadensis, but misidentified (and annotated) the second as Phaca neglecta (= A. neglectus). The first, properly considered the holotypus, was apparently collected by Kalm in Canada. The second is a cultivated plant from the Upsala garden and much resembles the diffuse type of A. canadensis, mentioned further below, which is preserved as A. canadensis flore viridi flavescente in the herbarium of Dillenius.

The identity of A. carolinianus is somewhat controversial, for no type-specimen appears to exist. The account of the species in Species Plantarum is based entirely on Dillenius (Hort. Eltham. 1: 45, Tab. XXXIX. 1732), for the synonym of Royen (Fl. Leidensis 391. 1740) is merely a quotation from the same source. Even the key character differentiating between A. carolinanus and A. canadensis (stems respectively erect and diffuse) was borrowed by Linnaeus from Dillenius, and we must turn to Hortus Elthamensis and the Dillenian herbaria for firsthand information. Dillenius described two American Astragali which he recognized as related: an A. Canadensis flore e viridi flavescenti identified, probably correctly, with the plant published earlier by Dodart under the same phrase-name; and a new A. procerior non repens flore e viridi flavescenti, grown from seed collected in Carolina, the basis of the Linnaean A. carolinianus. Dillenius brought out several seemingly important differences between his two species. His A. Canadensis etc. was a comparatively low and diffuse plant from a creeping root, seldom over two foot (6 dm.) tall, with weak branching stems, 7-10 pairs of subglabrous leaflets pallid-glaucescent but not pubescent beneath, and flowers yellowish-green on opening but turning a dull sooty color. By contrast A. procerior etc. had several stout stems up to 2-3 cubits (± 9-13.5 dm.) tall arising together from the same root-crown, and leaflets in 7 or 8 up to 20 pairs, green and glabrous on the upper side and paler but also pubescent beneath. Dillenius insisted, however, that the chief differential character was found in the root, saying that A. procerior etc. "radicibus non repit," whereas "ille [A. Canadensis etc.] vero insigniter." A creeping subterranean caudex is an important feature of sect. Uliginosi.

The Sherard herbarium (OXF) possesses two specimens ostensibly from the Dillenian collections, which' have a bearing on the identity of these two astragali, but they are far from conclusive and raise even more problems than they resolve. The first is labeled "A. Canadensis flore viridi flavescenti Dodart" and agrees pretty well with the description in Hortus Elthamensis. It represents an unusually slender, apparently sprawling form of var. canadensis; but as it consists of only part of the stem, the full stature cannot be determined and the nature of the root is, of course, unknown. The leaflets number 11-19, agreeing with the Dillenian diagnosis, but the pubescence is not really in accord. The specimen is annotated in an old (unfortunately unidentified) hand with the remark: "videtur potius procerior, non repens," but it conforms poorly with the original description of A. procerior etc. The second Dillenian specimen is a Catesby collection from Carolina, dated 1724. According to Vines & Druce (The Dillenian Herbaria 160. 1907) this ought to represent A. procerior non repens of Hortus Elthamensis, i.e., A. carolinianus L. Actually it is labeled "A. Carolinianus foliis angustioribus glabris, floribus pallide luteis in spicis longissimis laxius dispositis," does not agree at all closely with the description of A. procerior etc., and represents a quite different species, the native Carolinian A. Michauxii. So far as I was able to determine in 1962, no specimen of the genuine A. procerior etc. has survived among the Dillenian collections; and since there is no material of A. carolinianus in the Linnean herbarium, it seems proper to select Plate 39 of Hortus Elthamensis, which Linnaeus cited, as typus of the latter.

Several questions about the source and identity of A. procerior etc., and hence of A. carolinianus, remain unanswered. The Dillenian plate illustrates very exactly a robust form of A. canadensis, which might have originated in the Mississippi Valley but is not known to exist either in North or South Carolina today. It is definitely not the Appalachian form of var. canadensis which occurs in the Carolinas and which has usurped the epithet carolinianus. A robust form of var. canadensis does occur along the James River near Williamsburg in southeastern Virginia and might have been introduced from that region into English gardens. But if the A. procerior etc. described by Dillenius was really a form of A. canadensis, it is most difficult to explain the differential character of a non-creeping root on which so much emphasis was placed. Curiously enough, the feature of a determinate (non-creeping) root-stock, accompanied by tall, wandlike stems and many pairs of leaflets, is characteristic of A. Michauxii which had been collected in fruit (and with apparently ripe seed) in Carolina and communicated to Dillenius about 1724. But the leaves of A. Michauxii are not pubescent beneath, the flowers are purplish and not greenish-yellow, and the raceme is much looser than that of any form of A. canadensis. Moreover, the floral details, especially the long wing-auricles, shown in the plate of A. procerior etc., are definitely taken from A. canadensis, not A. Michauxii. We may conjecture that Dillenius did actually see A. Michauxii growing and derived the root character and geographic origin of his A. procerior etc. from that source; but later through some unexplained confusion of materials, he associated these details with a description and illustration of an altogether different plant, a form of A. canadensis probably not native to Carolina, North or South. The assumption that A. carolinianus came from the mountains of interior Carolina dates from Michaux’s Flora Boreali-Americana (1802).

Distribution

British Columbia Canada North America| Alberta Canada North America| Saskatchewan Canada North America| Manitoba Canada North America| Ontario Canada North America| Québec Canada North America| Northwest Territories Canada North America| Washington United States of America North America| Idaho United States of America North America| Montana United States of America North America| Wyoming United States of America North America| Colorado United States of America North America| Utah United States of America North America| New Mexico United States of America North America| Nebraska United States of America North America| Kansas United States of America North America| North Dakota United States of America North America| Oklahoma United States of America North America| South Dakota United States of America North America| Texas United States of America North America| Louisiana United States of America North America| Arkansas United States of America North America| Missouri United States of America North America| Iowa United States of America North America| Minnesota United States of America North America| Wisconsin United States of America North America| Illinois United States of America North America| Michigan United States of America North America| Ohio United States of America North America| Indiana United States of America North America| Kentucky United States of America North America| Alabama United States of America North America| North Carolina United States of America North America| South Carolina United States of America North America| Georgia United States of America North America| Virginia United States of America North America| West Virginia United States of America North America| Pennsylvania United States of America North America| New Jersey United States of America North America| New York United States of America North America| Vermont United States of America North America|