key: cord-022643-2j559muh
authors: Chance, Gail D.; Bacon, John D.
title: SYSTEMATIC IMPLICATIONS OF SEED COAT MORPHOLOGY IN NAMA (HYDROPHYLLACEAE)
date: 1984-07-01
journal: Am J Bot
DOI: 10.1002/j.1537-2197.1984.tb14148.x
sha: 
doc_id: 22643
cord_uid: 2j559muh

Seeds from 37 species of Nama (Hydrophyllaceae) were examined by scanning electron microscopy in order to assess systematic implications of seed coat structure. Generally, seed coat morphology is species specific. Nevertheless, similarities among species in seed coat organization, particularly, outer testa anatomy and wall thickening‐pitting patterns allow the recognition of six groups among examined taxa; furthermore, seed coat features often suggest both intra‐group and inter‐group relationships. Recognized groupings do not correspond well with the more classicial treatments of Nama; rather, they suggest that a restructuring of the infrageneric and even the generic taxonomy of Nama is in order.

THE SECOND LARGEST GENUS in the Hydrophyllaceae, Nama is a predominantly North American taxon common to more xeric regions of northern Mexico and southwestern United States. However, a few species extend into South America and, curiously enough, one is endemic to the Hawaiian Islands. While the genus, on the whole, is widespread, most species are limited in distribution often by edaphic factors (e.g., gypsum, Johnston, 1941; Bacon, 1981) .

The excellent classical treatment of Nama provided by C. L. Hitchcock (1933 Hitchcock ( , 1939 continues as the major source of information regarding the genus. Hitchcock recognized 40 species, ranging from delicate annuals to robust, woody perennials. These were distributed among five sections: Arachnoidea (1 sp.), 0nerascentia (1 sp.), Zonolacus (1 sp.), Conanthus (2 spp.), and Nama (as Eunama, 35 spp.). Subsequently, additional species have been recognized, one in sect. Conanthus and the remainder in sect. Nama, so thattheir number now totals about 50. Each of Hitchcock's smaller sections is well defined and segregated from others by singularly notable characters. Indeed, sects. Arachnoidea and Cinerascentia are so well defined that their placement in Nama has been questioned (Bacon, 1974; Raven and Axelrod, 1978) . In contrast, sect. Nama houses a diverse assemblage of species, as noted by Hitchcock (1933) , and relationships among constituent taxa, as well as intersectional relationships, have yet to be fully explored. Hitchcock's (1933) observation that seeds of different species exhibit variously patterned surfaces offers a promising point of departure for studies aimed at elucidating relationships in Nama. Indeed, seed coats may retain ingenuous indications of phyletic commonality (Martin, 1946; Davis and Heywood, 1963; Esau, 1977) , and many recent studies in various, unrelated groups have confirmed the general utility of seed coat characters for assessing systematic relationships (e.g., Chuang and Heckard, 1972; Whiffin and Tomb, 1972; Hill, 1976; Clark and Jernstedt, 1978; Canne, 1979 Canne, , 1980 Elisens and Tomb, 1983) . We have examined the seed coats of 37 species of Nama utilizing scanning electron microscopy (SEM) and report implications of seed coat structure towards systematics of Nama.

MATERIALS AND METHODS -Mature seeds removed from herbarium specimens were mounted on brass or aluminum specimen stubs with double-stick Scotch tape. Specimens were then coated with gold or gold-palladium in either a Technics Hummer Jr. or a Polaron E-5100 vacuum sputter coater. Sputter coaters were calibrated so that coating for 2.5 minutes yielded a metal layer of about 200 Ain thickness.

In order to examine the fine features of the testa and its component cells, seeds were sectioned free-hand with razor blades. Sections were extracted in acetone for 30-60 sec to remove lipid material present in the endosperm layer which, on sectioning, covered the testa layer and obscured structural detail (Amott and Webb, 1983) . Sections were mounted and coated as described for whole seed mounts. [Vol. 71 Additionally, some seeds were soaked for 30-60 min, or longer ifnecessary, in distilled water or a 1% solution of Aerosol OT to facilitate removal of seed coat pieces, peels, from the whole seed. Peels were mounted on specimen stubs and sputter coated as previously described.

RESULTS AND DISCUSSION-Seeds of Nama species are morphologically diverse. Ranging from 0.3 mm to 1.5 mm in length, seeds are black, brown, yellow or white in color and are spherical, oval, fusiform, ovate or irregularly multi-sided in shape. However, as noted in other plant groups (Chuang and Heckard, 1972; Tomb, 1974; Clark and Jernstedt, 1978; Canne, 1980) , the shape and size of mature seeds may vary in some Nama species. Variation appears to be primarily related to the number and position of ovules developing within the ovary (Amott, 1962; Tomb, 1974) . Nevertheless, for each species of Nama there is a typical shape and size ( Fig. 1-62) .

Viewed in section, mature seeds possess an outer testa that is either solid, thus appearing as a dense, melded, acellular layer, or else chambered (appearing as a layer clearly cellular in composition). In surface patterning, seeds with a solid coat generally appear foveolate, papillose or exhibit gentle depressions. Seeds with a chambered coat are moderately to strongly reticulate with the degree of reticulation being dependent upon the relative wall height of cells forming the reticulum. Moreover, the reticulum provides the most discriminating features of the chambered seed coat, namely thickenings or pits-perforations in the walls of reticulum chambers.

Utilizing seed coat features, seeds of examined taxa may be divided into six distinct groups. Within each group seed size, shape and coat particulars may further characterize each species. We choose not to describe each species; rather, micrographs are included to convey the essence of species individuality. Descriptions are restricted to the minimum necessary to facilitate our discussion and systematic appraisal of each group.

Seed Group 1 (Fig. 1-4 )-The two included species, N. rothrockii (Fig. 1, 3) and N. lobbii (Fig. 2,4) , produce seeds with a thick solid coat 45-60~m in thickness. Moreover, the seeds of both species average 1.4 mm in length and are the largest of those examined.

For both species we infer that the outer testa begins as a cellular layer and develops into a solid layer after intracellular deposition ofwall material. The porous, honeycombed appear-ance of the upper Ih~h of the outer testa in N. rothrockii (Fig. 3) as well as the small gaps associated with papillose extensions (suggestive of cell junctions) in N. lobbii (Fig. 4) support this contention. We attribute surface patterning of the two to collapse of the upper tangential wall to the level of deposited material. Radial walls in N. rothrockii are essentially the same height; collapse of the upper wall results in a foveolate surface. Certain areas of the radial walls in N. lobbii are elongated and higher than the remaining wall; collapse of the upper tangential wall produces a papillose surface with each papillus being a wall extension. Hitchcock (1933) segregated N. lobbii and N. rothrockii into the monotypic sections Arachnoidea and Cinerascentia, respectively; the former species produces capsules which are loculicidally and septicidally dehiscent while other namas exhibit only loculicidal dehiscence. Nama rothrockii is the sole Nama with capitate inflorescences. Moreover, chromosome numbers have been determined for well over 3f4 ofthe species ofNama (Cave and Constance, 1947 , 1950 , 1959 Constance, 1963; Bacon, 1974) , and establish the genus as a strikingly diploid group with a base number of x = 7. N. lobbii is known only as tetraploid, n = 14, while N. rothrockii is even more radically divergent with n = 17. Seeds ofthe two species also diverge from the bulk of Nama. They are unique in terms ofsurface patterning and thickness of the outer testa layer among examined species. Indeed, features ofsize and coat thickness would argue that the two are more closely related one to another than either is to yet other namas. We perceive the former relationship as distant. Clearly, seed features support our opinion that inclusion of these two species in Nama is less than tenable.

Seed Group 2 (Fig. 5-10 )-This seed group comprises five species characterized by a thin, solid seed coat 2-5~m in thickness. Surface patterning is generally foveolate. The foveae are not formed because of unequal thickening of the seed coat, since the coat is essentially uniform in thickness. Rather, the coat conforms to the underlying endosperm topology resulting in shallow surface depressions (Fig. 8, 10) .

Seeds of four included species, N. hispidum (Fig. 5) , N. turneri, N. stevensii, and N. undulatum (Fig. 6 ), reveal surface striations as well as ridges believed to reflect cell boundaries (Fig. 7) . The fifth species, N. sandwicense, exhibits a bullate surface (Fig. 10) . Additionally, seeds of the former four species are yellow in color and fusiform in outline, while those of the latter taxon are generally gray-white in color and ellipsoid.

Similarities among the seeds ofN hispidum, N. turneri, N stevensii, and N. undulatum suggest the four are closely related; indeed, a relationship between N hispidum and N turneri has been proposed (Bacon, 1981) . While seed features ofthe anomalous N sandwicense suggest this Hawaiian endemic is distantly related, similarities previously enumerated argue that it is allied to this group.

Seeds of species in Group 2 are easily distinguished from those in Group 1, but the fact that both groups possess solid coats hints at a common origin. However, despite external appearances, we have observed no indications suggestive of a cellular structure in seed coat sections among species of Group 2. It is possible, therefore, that the surface ridges arise due to compression of the ovule during development and that the coat may be formed partly from an extracellular exudate rather than wholly from an intracellular deposit as we infer for the species of Group 1. In any event, it seems more plausible to view the solid seed coats of Group 1 and Group 2 not as characters based in a common origin, but as characters independently acquired in only remotely related groups.

Seed Group 3 (Fig.11-19 )-Thesixincluded species are characterized by a reticulate seed coat in which pits-perforations (Fig. 11 , 15) are found in radial walls of the reticulum cells. Additionally, the apical rim of each chamber is "knobby" in appearance (Fig. 18, 19 ), although on individual seeds the knobs may be obscured by the adherent tangential wall. Generally, seeds ofthis group are ovoid and brown in color.

Seeds ofNjamaicense (Fig. 13 ) are identical with those of N bart/ettii. In both, reticulum cells are arranged into definite rows; each cell is elongated at right angles to the long axis of the seed and the longer radial walls are conspicuously concave (Fig. 13 ). This clearly contradicts the suggestion of Standley (1940) that N. bartlettii is most closely related to N. undulatum, for the latter resides in Seed Group 2.

Seeds ofN. marshii and N. propinquum ( Fig.  12 ) also are identical. In contrast to N jamaicense and N bartlettii, reticulum cells of these are more irregular in shape, size and arrangement; some radial walls may be weakly concave (Fig. 12) . Seed features ofthe two confirm the relationship perceived by Johnston (1943) .

Much variation in seed features occurs in N palmeri. Extreme seed forms include a larger Fig. 1-10 [Vol. 71 morph (Fig. 17) in which reticulum cells are irregularly ellipsoid and exhibit prominent rim knobs (Fig. 18 ) and a smaller morph (Fig. 16 ) with more rounded, somewhat angulate chambers and less prominent knobs (Fig. 19) . Connecting the two extremes is a continuum of seed forms. Variation in seed features is in keeping with variation in other morphological features as noted by Hitchcock (1933) , who treated the species as consisting of two poorly defined, intergrading varieties.

Emphasizing its half-inferior ovary, Hitchcock (1933) segregated N. stenocarpum (Fig.  14) into the monotypic sect. Zonolacus. Seed features suggest, however, that the species is to be more closely allied with the taxa ofGroup 3 than its present placement would indicate. Indeed, seed features argue that the half-inferior ovary is more ofa "technical" than a "phyletic" character. It is noteworthy, too, that at maturity both N. palmeri and N. jamaicense exhibit hardened calyx lobes which tightly invest the capsule; and, while the calyx lobes are never adnate to the capsule, mature calyx-capsule units are reminiscent of the fruits of N. stenocarpum.

Seed Group 4 (Fig. 20-35 )-Each ofthe nine species in this group produces a chambered coat ranging from weakly to deeply reticulate; and all possess localized thickenings in the radial walls of reticulum cells (Fig. 22, 23, 27-29, 32, 35) . Moreover, each species produces characteristic, easily distinguishable seeds; yet, reticulum organization suggests even more intimate relationships among certain species.

Seeds of N. stenophyllum (Fig. 20, 22 ) and N. canescens (Fig. 21, 23 ) exhibit a deep reticulum with elaborate thickenings which are typically reticulate in the former and reticulate-sca1ariform in the latter. These features clearly support the relationship proposed by Hitchcock (1939) .

Certain cells in the reticulum ofN. johnstonii undergo asymmetrical wall development so that one portion of their wall is greatly elongated and projects outward along the largest circumference of the seed (Fig. 30) ; remaining cells have walls of equal height and are moderately reticulate. Less pronounced asymmetrical wall development is also exhibited by seeds of N. hitchcockii (Fig. 26) . Thickenings in all cells of the latter species and symmetrical cells of the former are generally columnar and rarely branched (Fig. 29) ; in asymmetrical wall extensions of N. johnstonii thickenings are sealariform-reticulate (Fig. 30) . Furthermore, seeds of N. jlavescens (Fig. 33) , while lacking asymmetrical walls, exhibit reticulum cells similar to the symmetrical cells ofthe previous species. Wall thickenings in this species are also columnar and unbranched. Seeds of N. havardii (Fig. 31 ) are similar in shape and sizeto those of N. jlavescens and are characterized by columnar wall thickenings (Fig. 32) . However, thickenings in N. havardii are more stout and evenly spaced than in any of the previous species.

Both N. carnosum (Fig. 24) and N. constancei (Fig. 25) produce seeds characterized by columnar wall thickenings. In contrast to other Group 4 species, however, thickenings in these species are not apically interconnected; that is, they lack a continuous thickening circumscribing the upper cell margin (Fig. 27, 28) . As a result, the reticulum in these species is characteristically knobby, each knob representing the apex of a thickening.

Nama prostratum (Fig. 34) produces one of the more striking seed coats in the genus. Reticulum cells exhibit asymmetrical wall development and the taller wall portion is elaborately patterned with scalariform-reticulate thickenings (Fig. 35) . In contrast to other species ofits group, unbranched thickenings often traverse the floor of reticulum cells (Fig. 35) and are reminiscent of those found in Seed Group 5.

Of the nine species included in this group, Hitchcock (1933 Hitchcock ( , 1939 was aware of seven and considered five (N. stenophyllum, N. canescens, N. jlavescens, N. johnstonii, and N. carnosum) as closely related (Hitchcock, 1939) . The two unknown to him, N. hitchcockii and N. constancei, are recently described (Bacon, 1981) . All seven are woody-based perennials with sessile, revolute, essentially linear leaves. All are Chihuahuan Desert endemics, with the exception of N. hitchcockii, found on drier, western slopes of the mountains forming the southeastern margin of the desert. Reticulum organization further attests to the unity of this "linear-leaved" grouping and is even more instructive as to relationships among them.

The suggested alliance of N. havardii to aforementioned species is not surprising since it, too, is a Chihuahuan Desert endemic. Seed features indicate that N. havardii is a more distant relative which is consistent with its herbaceous, robust habit and its petiolate, broader, generally revolute leaves.

Nama prostratum departs markedly from other members of Group 4 since it is a weakstemmed perennial found in the mesic highlands of west-central, central and southern Mexico. Its distinctive seed coat and habit suggests early divergence from ancestral stock out of which arose the desert species to the north. Indeed, this taxon may prove to be representative of a stock out of which Seed Groups 3, 4 and 5 all arose since, in many ways, seeds of N. prostratum are archetypical of seeds in all three groups. Significantly, branched thickenings are limited to the more vertically elongated portions of reticulum walls in N. prostratum as they are in N. johnstonii (they are found elsewhere only in the deeply reticulate coats of N. canescens and N. stenophyllum), while unbranched thickenings are found in the shorter wall portion, as they are found in the less deeply reticulate, symmetrical chambers of other Group 4 species. Moreover, unbranched thickenings commonly present on chamber floors of N. prostratum seeds often continue into the shorter wall of the chamber, much as those thickenings characterizing Group 5 species; all Group 5 species exhibit a shallower reticulum. Additionally, we suggest that concave reticulum walls in Group 3 species are merely an alternative manifestation of asymmetrical wall development. And, since thickenings may differ from pits only in degree, widening ofreticulate thickenings in N. prostratum could yield pits, as found in species of Group 3. However, we believe that N. prostratum is correctly housed, although well removed from other Group 4 species.

Seed Group 5 (Fig. 36-44 )-Seeds ofthe six species in this group are moderately reticulate and exhibit many "u-shaped" thickenings in each reticulum chamber (Fig. 36) . Each thickening is continuous, from one wall across the chamber floor and into the adjacent wall; apices of thickenings, on collapse of the upper tangential wall, produce the characteristic knobs along the upper margin of each cell. Each of the included species is quite distinct morphologically and differs from the others in habit, leaf shape and size, pubescence and corolla shape, size and color. Hitchcock (1933) noted the similarities between the seeds of N. torynophyllum (Fig. 39) and N. parvifolium (Fig. 41) . It was his opinion, however, that N. parvifolium was most closely related to N. prostratum and, to a lesser degree, apparently, to If. rotundifolium. We have pointed out the 'potential relationshp of N. prostratum to Seed group 5, including N. parvifolium. We view it as remote. Seed features confirm a kinship between N. parvifolium and N. rotundifolium and, as Hitchcock implied, suggest it is not direct. For, were it not for its wider reticulum cells, seeds ofN. rotundifolium (Fig.40) would be essentially identical to those of N. torynophyllum (Fig. 39) . And, although producing well developed thickenings and a prominently knobbed reticulum as do the former species, seeds of N. parvifolium (Fig. 41) are much like those of N. serpylloides (Fig. 42) and N. rzedowskii (Fig. 43, 44 ). In these three species, reticulum cells are less elongated, broader and more irregularly arranged than in N. rotundifolium and N. torynophyllum. Seed features argue that N. serpylloides,N. rzedowskii and N. parvifolium are a related trio.

Nama hirsutum (Fig. 38) is difficult to relate directly to other members ofGroup 5. It occurs in the mesic, but seasonally dry, highlands of southern Mexico and Guatemala; whereas, other Group 5 members are inhabitants of the drier desert regions to the north. Its reticulum cells exhibit clearly asymmetrical walls and in shape resemble those in Group 3. But, thickenings characteristic of Group 5 species are present in its reticulum cells (Fig. 37 ) and countenance its placement. Nevertheless, similarity of the seeds of N. hirsutum to those of Group 3 strengthen even more our previous postulations of a common stock from which Group 3, 4 and 5 arose; for its seeds recall, too, in their reticulum asymmetry, those of N. prostratum. In short, we view the more mesic southern distribution ofthese two taxa as more than fortuitous and suggest that N. hirsutum likely represents yet another early offshoot of that stock.

Seed Group 6 ( Fig. 45--62 )-Chamberedseeds with undulate chamber walls characterize the species of this group. However, seed features are more diverse in the group than in the others established and we shoulder some misgivings in our structuring of this group.

Our major concern relates to our placement of N. origanifolium (Fig. 45) and N. sericeum (Fig. 48) . In many respects, seeds of N. origanifolium are reminiscent of those of Group 5. Indeed, in surface view wall undulations often suggest the "u-shaped" thickenings characteristic of the former group. However, in neither sections nor peels have we observed such thickenings and we attribute the resemblance to folds in the collapsed upper tangential waD where the wall encounters more prominent undulations. In fact, sections and peels reveal chamber walls to be characterized by pores in this taxon (Fig. 46 ) as well as in N. sericeum. With its irregular reticulum organization, N. sericeum (Fig. 48 ) also is reminiscent of seeds of certain species in Groups 3, 4 and 5. We are aware that the pores must be related to thickening-pitting alternatives. Further study may well prove these two species to have closer relatives elsewhere, but they are included herein due to their undulate walls and [Vol. 71 because four of the other seven species in this group exhibit similar pores in their chamber walls.

In contrast to the woody-based, perennial habit ofN. sericeum and N. origanifolium, other members of Group 6 are annual and strikingly similar in their dichotomously branched growth habit. While reticulum cells in seeds of N. demissum (Fig. 47) and N. dichotomum (Fig.  49 ) tend to resemble more those of N. origanifolium, seeds ofthe former two species exhibit the weak depressions characteristic of seeds of yet other species suggesting that their placement is correct.

In contrast to other examined species, seeds of N. aretioides (Fig. 50, 51) , N. densum (Fig.  60, 61) , N. parviflorum (Fig. 58, 59 ) and N. depressum (Fig. 55, 57 ) exhibit dimorphic testa patterns. One morph exhibits a weakly chambered testa (Fig. 51 , 57, 58, 61), while the alternate form lacks distinct chambers and is characterized by minute pits-pores on the seed surface ( Fig. 55, 59, 60 ). Internal walls of the chambered morph of these species are characterized by pores (Fig. 53) as are the nonchambered form of N. densum and N. parvijlorum (Fig. 54) . In contrast, sections of the non-chambered forms of N. aretioides and N. depressum reveal uniformly dense areas, presumably lateral walls, enclosing a porous-granular material, presumably occupying the cell lumen, and pores have yet to be seen in these forms.

In N. aretioides the two morphs are clearly distinct in surface patterning (Fig. 50, 51) ; whereas, in N. densum (Fig. 60, 61) , N. parviflorum (Fig. 58, 59 ) and N. depressum (Fig. 55, 57) differences between the two morphs are more subtle. Indeed, the non-chambered morphs of the latter three species exhibit surface ridges which are strongly reminiscent of lateral walls. The resemblance is accentuated particularly in some seeds of N. densum and N. parviflorum by restriction of surface pits to such ridges (Fig. 56) . In contrast, non-chambered seeds of N. aretioides lack surface ridges and pits are more randomly distributed. Individual reticulum cells in the chambered form ofN. aretioides differ in shape and organization as compared with the chambered morph ofthe other three species; and, contiguous cell walls are discernablein N. aretioides (Fig. 52) while they are not apparent in the others.

The seed morphs exhibited by N. densum, N. depressum and N. parviflorum are similar enough to suggest that the non-chambered morph might be merely immature. Moreover, the chambered morphs ofthe three species are very similar among themselves and one might expect a similar pattern of development for their seed coats. Such a similarity could lead to similar immature seed forms. While the seed morphs produced by N. aretioides are easily distinguished from each other and from seeds of the previous species, its non-chambered morph could also be an immature form. It seems unusual, however, that both chambered and non-chambered morphs were never encountered in the same collection ofany species. If developmental age were responsible for the dimorphism, one might expect to encounter occasionally seeds in various stages of development.

On the other hand, the dimorphism might reflect genetic differences since examined collections of each species were geographically separated. Hitchcock (1933) notes that N. aretioides from Washington, Idaho and southwestern California has smaller corollas and wider leaves than cohorts from Nevada and central California. Furthermore, differences in ploidy level have been shown to correlate with different seed coat morphology in Portulaca oleracea (Danin, Baker and Baker, 1978) Chenopodium album (Wilson, 1980) , and Castilleja (Heckard, Morris and Chuang, 1980) ; N. parviflorum has been reported as both diploid (n = 7) and tetraploid (n = 14) (Constance, 1963) . The remaining species have been reported only as diploid, but reports are few in number. We are unable to explain the dimorphisms at present.

The final species comprised in Group 6, N. pusillum, exhibits a somewhat atypical reticulum, compared to other members, for the walls are only weakly, and not uniformly, undulate (Fig. 62) . Nevertheless, in section its internal walls exhibit the pores found in other species ofthis group; and, its weak surface depressions are in keeping with those exhibited by seeds of other members.

In spite of our misgivings as to the ultimate placement of N. origanifolium and N. sericeum, we are convinced that Seed Group 6 is essentially a natural alliance. Minimally, seed features would argue that N. aretioides, N. densum, N. parviflorum, N. depressum, and N. pusilium must be positioned closely one to another; the latter two species are housed in sect. Nama while the former three, because oftheir connate styles, constitute sect. Conanthus (Hitchcock, 1933) . Clearly, seed features suggest a more direct relationship of those species possessed of connate styles to yet other species with free styles than is presently implied. It is doubtful that the sect. Conanthus can be maintained.

CONCLUSIONs-Our SEM examination of Nama seeds confirms the general value of seed coat features as systematic indicators. Testa features clearly partition examined species into distinctive groups. Moreover, they often offer even further suggestions of intra-as well as inter-group associations. We view the species so grouped as phyletically based lineages. Therefore, a rearrangement ofthe infrageneric groupings within Nama is indicated. Nevertheless, we suggest no formal restructuring at present, preferring to await further confirmation of our views in the seed coats of as yet unexamined species.

The seed, germination, and seedling of Yucca

Scanning electron microscopy and histochemistry of oilseeds

New species of Nama (Hydrophyllaceae) from the Chihuahuan Desert region of Mexico

A light and scanning electron microscope study of seed morphology in Agalinis (Scrophulariaceae) and its taxonomic significance

Chromosome numbers in the Hydrophyllaceae

Seed coat morphology in Cordylanthus (Scrophulariaceae) and its taxonomic significance

Systematic studies of Eschscholzia (Papaveraceae). II. Seed coat microsculpturing

Chromosome number and classification in Hydrophyllaceae

Cytogeography and taxonomy ofthe Portulaca oleracea L. polyploid complex

Principles of angiosperm taxonomy

Seed morphology in New World Antirrhineae (Scrophulariaceae): systematic and phylogenetic implications

Plant anatomy

Origin and taxonomy of Castilleja montigena (Scrophulariaceae)

Taxonomic and phylogenetic significance ofseed coat microsculpturing in Mentzelia (Loasaceae) in Wyoming and adjacent states

A taxonomic study of the genus Nama

Gypsophily among Mexican desert plants

The comparative internal morphology ofseeds

Origin and relationships ofthe California flora

Studies of American Plants. XI. Field Mus

SEM studies of small seeds

The systematic significance of seed morphology in the neotropical capsular-fruited Melastomataceae

Artificial hybridization among species of Chenopodium sect

ApPENDIX: SOURCE OF SEED SAMPLES Collections indicated with an asterisk are the source of included photographs. Collections without herbarium designation will be deposited at TEX. Nama sect

Hartman 6400 (TEX)

Lemmon & wife

Ni flavescens Brandeg

N havardii A. Gray: CHIHUAHUA, Bacon 1042

Conzatti 191 (GH). N hispidum A. Gray: CHIHUAHUA, Bacon 963, 1039, 1045

Palmer 158 (GH). N palmeri A. Gray ex Hemsl

Bacon 1132*, 1490

Johnston etal. 11172(fEX)

N serpylloides A. Gray ex Hemsl. COAHUILA, Bacon 1318*

Hartman and Funk 4012 (TEX). N. torynophyllum Greenm

Chance et al. 10. N turneri Bacon