Cavendishia crassifolia (Benth.) Hemsl.
-
Authority
Luteyn, James L. 1976. A revision of the Mexican Central American species of Cavendishia (Vacciniaceae). Mem. New York Bot. Gard. 28 (3): 1-138.
-
Family
Ericaceae
-
Scientific Name
-
Description
Species Description - Epiphytic or terrestrial shrubs, 1-3 m tall, rarely arborescent and then to 10 m tall with stem bases to 20 cm in diameter. Mature branches terete or subterete, smooth or striate, glabrous; bark brown to grayish-brown; immature branches and twigs of n e w growth subterete or bluntly angled, striate or ribbed, glabrous to densely pubescent, green or often red above and green beneath. Petioles subterete, usually flattened adaxially, rugose, (3-)5-l I(-17) mm long and 1-4 mm in diameter, glabrous, puberulent or pilose, reddish-brown when dry. Leaves oblong, elliptic, ovate or lanceolate, rarely oblanceolate or semiorbicular, 4-15(-22) cm long and 1.5-5(-11) cm broad, basally rounded, cuneate, subcordate, subtruncate or rarely short attenuate, apically acute to acuminate, often abruptly, rarely rounded or cuspidate, green, olive-green, or brownish, sometimes metallic- or slate-blue when dry, glabrous or short pilose along veins above and below; (3-)5-7(-9)-plinerved, innermost lateral nerves sometimes rising 1-2 cm above base, midrib impressed above and raised beneath, always very conspicuous sometimes flattened adaxially, lateral nerves impressed or slightly raised distally above, raised beneath, veinlets impressed or raised above, slightly raised but usually obscure beneath. Inflorescence (4-)6-20 (-40) flowered, encircled at base by bracts which may be glabrous to densely short pubescent, or often pubescence restricted to margins or apex; rachis subterete to bluntly angled, (0.5-)1-5(-8) cm long, green or red, glabrous or pubescent with short white trichomes. Floral bracts smooth or muricate, glabrous to densely short pilose, sometimes pubescent only along margins or at distal tip, oblong, ovate, elliptic to oblanceolate, basally narrowed and truncate, apically rounded or acute, (11-)17-26(-40) mm long and (5-)10-17(-27) mm broad, pink to dark red sometimes provided with reddish, glandular fimbriae abaxially. Pedicels often swollen distally, striate or ribbed, sometimes rugose, (3-)6-I5(-20) mm long and 1-3 mm in diameter, pale green to dark rosered, glabrous or pilose, often with glandular fimbriae scattered along length or densely aggregated distally. Bracteoles oblong, ovate, lanceolate, linear-lanceolate, rarely aristate, 1-4(-8) mm long and 0.5-2(-5) mm broad, glabrous to densely glandular-fimbriate, sometimes cUiate. Calyx glabrous to densely short pilose, (3.5-)4-6(-10.5) mm long, pale green to red, usually with few to m any short, stout, glandular fimbriae; tube cyHndric, smooth or somewhat rugose, often somewhat pentagonal in cross-section, 1.5-3(-4) mm long, nonapophysate, base often rounded or truncate and shallowly undulate; limb cylindric to campanulate, usually somewhat spreading, (1.5-)2-3(-6) mm long including lobes; lobes triangular, rarely oblong, (0.5-)l-2(-5) mm long, connivent after anthesis, marginally glandular-fimbriate; sinus broadly rounded or acute. Corolla cylindric to bottle-shaped, narrowed to throat, (12-) 15-20(-25) mm long and 4-5 mm in diameter, glabrous to short pilose, with or without scattered glandular fimbriae, pale pinkish to white at base, otherwise tube dark pink to red, with limb and lobes white; lobes deltoid, widely flaring at anthesis, 1-2 mm long. Stamens 11.5-19 mm long; filaments slightiy coherent at base, but more commonly distinct, glabrous or distaHy pubescent, alternately either 2-4 mm or 3.5-6.5 mm long, white or pinkish-white; anthers including tubules alternately either 7.5-14.5 mm or 10-16.5 mm long, yellow to orange; thecae 2.5-6 mm long. Style glabrous or rarely distally short pUose, (12-)14-19(-27) mm long. Berries 8-14 mm in diameter. Phenology: Flowering and fruiting throughout the year, very local.
-
Discussion
Thibaudia crassifolia Bentham, Pl. Hartw. 65. 1840. Type: Mexico: State of Oaxaca. Totontepeque, Hartweg 477(Holotype: K! photo of type DUKE! isotypes: G! NY! US-fragments! W!).
Thibaudia mexicana Martens & Galeotti, Bull. Acad. Sci. Brux. 9(2): 530. 1842. Type: Mexico: State of Oaxaca. Chinantle, Galeotti 1813 (Holotype: BR! photo of type DUKE!).Proclesia veraguensLs Klotzsch, Linnaea 24: 35. 1851. Type: Costa Rica. Warszewicz s.n. (Holotype: B, not seen, presumably destroyed, photo of type DUKE! NY!).
Polyboea crassifolia (Benth.) Klotzsch, Linnaea 24: 31. 1851.Cavendishia latifolia Hemsley, Biol. Centr. Am. Bot. 2: 273. 1881. Type: Mexico: State of Chiapas. Pueblo Nuevo, Linden 390 (Holotype: K! photo of type DUKE! isotypes: F-fragments! G!).
Cavendishia veraguensis (Kl.) Hemsley, Biol. Centr. Am. Bot. 2: 273. 1881.
Chupalon crassifolium (Benth.) Kuntze, Rev. Gen. Pl. 2: 383. 1891.
Chupalon latifolium (Hemsley) Kuntze, Rev. Gen. Pl. 2: 383. 1891.
Cavendishia guatemalensis Loesener, Bull. Herb. Boiss. II. 3: 221. 1903. Type: Guatemala: Department of Huehuetenango. Jacaltenango, Seler 3107 (Holotype: B, not seen, presumably destroyed, photo of type DUKE! NY! isotypes: GH! NY! US!).
Cavendishia smithii Hoerold, Bot. Jahrb. 42: 328. 1909. Type: Costa Rica: Province of Cartago. Volcan Irazii, Donnell-Smith 4876 (Holotype: B, not seen, presumably destroyed, photo of type DUKE! NY! isotypes: GH! US!).
Cavendishia costaricensis Hoerold, Bot. Jahrb. 42: 326. 1909. Type: Costa Rica: Province of San Jose. La Palma, Werckle 54 (Holotype: B, not seen, presumably destroyed, photo of type DUKE! NY! US! isotype: NY-fragments!).
Cavendishia hoffmannii Hoerold, Bot. Jahrb. 42: 328. 1909. Type: Costa Rica: Province of Cartago. Volcan Irazii, Hoffmann 141 (Holotype: B, not seen, presumably destroyed, photo and fragments of type NY! photo of type DUKE!).
Cavendishia chiapensis Brandegee, Univ. California Publ. Bot. 6: 188. 1915. Type: Mexico: State of Chiapas. Cerro del Boqueron, Purpus 7342 (Holotype: UC! photo of type DUKE! isotypes: F! GH! MO! NY! US!).
Cavendishia skutclui A. C. Smith, Jour. Wash. Acad. 27(7): 308-9. 1937. Type: Costa Rica: Province of San Jose. El General, Skutch 2802 (Holotype: US! photo of type DUKE! Isotypes: G H ! MICH! NY-fragments! S!).
Cavendishia calycina A. C. Smith, Ann. Missouri Bot. Gard. 28: 447. 1941. Type: Panama: Province of Chiriqui. Boquete, Davidson 157 (Holotype: A! photo of type DUKE!).
Cavendishia guatemalensis var chiapensis (Brand.) L. O. Williams, Fieldiana Bot. 31: 169. 1965.
Cavendishia crassifolia is the most frequently encountered species of Cavendishia in Mexico and Central America. It is here recognized as a complex of widely ranging, variable populations. Previous authorities have subdivided the complex into n o fewer than 11 taxa. All of these taxa were described from single collections, from different geographical areas, and were based upon characters which are n o w proven to be unstable. The complex also seems to range into South America where similar problems are to be expected. Future research into the Andean taxa will undoubtedly result in a reinterpretation of the number of recognized taxa.
Since a rather large number of well established names must be reduced to synonymy, some explanation is in order. Intensive field studies and the accumulation of numerous collections from throughout Mexico and Central America have shown that the morphological features which previously served to differentiate the taxa obviously blend together with n o clear-cut discontinuities. These populations and individual characters are discussed below in detaU. The populations which are found from Mexico to Honduras have previously been divided into four distinct but closely aUied species. The taxa have been characterized as follows (fide Smith, 1932):
1) C. crassifolia—glabrous, rachis 2-5 cm long, corolla 15 mm long, leaves 5-plinerved.
2) C. latifolia—glabrous, rachis 3-4 cm long, corolla 12-16 mm long, leaves 7-plinerved.
3) C. chiapensis—glsihrous, rachis 2-2.5 cm long, corolla 15-16 mm long, leaves 5-7-plinerved.
4) C. guatemalensis—infiorescence (including rachis, pedicels, bracts, bracteoles, calyx, and coroHa) pubescent, rachis 2-3.5 c m long, corolla 15 mm long, leaves 5-7-plinerved.
Hemsley's C. latifolia was characterized by its broadly ovate leaves with 7(-9) nerves. Three sheets of the type collection were available to me, and they differ from C. crassifolia only with respect to the wider leaves. Plants collected in Guatemala and Mexico show considerable variation in the size and shape of the leaves as well as in the number of nerves. Some plants may have elliptic to suborbicular leaves on either the same plant, or on plants growing next to each other. The number of primary nerves per leaf may also vary from 5-9 and seems independent of leaf width. In all other respects these plants differ very little from typical C. crassifolia. Cavendishia latifolia should only be considered as a wideleaved form of C. crassifolia, and the size and shape ofthe leaves, along with the number of primary nerves should not be used as criteria for specific delimitation. Considerable emphasis has also been placed upon rachis length. Cavendishia crassifolia was segregated from C. latifolia, C. chiapensis, and C. guatemalensis on the basis of its longer rachis which supposedly averaged 3 cm long or more, in contrast with the latter three taxa in which the rachis was less than 3 cm long. Both field observations and available specimens have shown that rachis length varies considerably across the range. Some populations may be very stable with respect to having either short or long rachises; however, other populations may show rachises of varying lengths. Not only may the mature rachises of a single plant vary in length, but it also seems that rachises may lengthen considerably after anthesis. Luteyn & Almeda 3425 (Oaxaca, Mexico) shows rachises from the lower portions of a branch measuring about 2.5 cm long, when those from the upper portions of the same branch measure 5-6.5 cm long. Therefore, rachis length should not be used as a criterion for specific delimitation. Pubescence patterns are also extremely variable. When first described, C. guatemalensis was distinguished by its pubescent rachises, bracts, pedicels, calyces, and corollas (Loesener, 1903). This combination of characters made C. guatemalensis a seemingly distinct entity. The numerous collections now available, however, show that every degree of pubescence m a y be demonstrated throughout the range, and therefore, the assignment of any taxonomic status on the basis of pubescence is unwarranted.
Both Smith (1932) and WiUiams (Standley and Williams, 1965) characterize C. crassifolia as having "bracts often shorter than the pedicels, not covering the inflorescence." However, their descriptions list pedicels 6-11 mm long and bracts up to 20 mm long. In m y observations, the pedicels are never longer than their subtending bracts, not even after anthesis. (This character of bracts shorter than pedicels was used by Smith in 1932 to single out C. laurifolia, C. quereme, and C. crassifolia. It holds for the first two but not for C. crassifolia, as is true of his other character of corolla length used in his 1932 key to species.
From Nicaragua to Panama the populations of C crassifolia have been divided into seven closely allied taxa, based primarily upon leaf characters and on pubescence.
The taxa have been characterized as follows (fide Smith, 1932):
1) C. smithii — leaves 2.2 times longer than broad, with abruptly caudateacuminate apices
2) C. veraguensis — pubescent calyx and corolla
3) C. costaricensis — narrow leaves (greater than 3 times longer than broad), with long acuminate apices
4) C. hoffmannii — leaves conspicuously punctate on both surfaces
5) C. skutchii — long rachises, metalic-colored leaves, subcordate leaf bases, short corollas
6) C. calycina—long calyx limb and lobes which are oblong and undulate margined, short pedicels and long bracteoles
7) C. miconioides — small, obloag, obtuselv short-acuminate leaves.
The most problematic taxa are the first three listed above, viz. C. smithii, C. veraguensis, and C. costaricensis. In 1932, Smith maintained the three taxa as distinct although he stated that they were "not too clearly separable." He also stated that "in floral characters the three species (viz. C. veraguensis, C. costaricensis, and C. hoffmannii) are very close, and I should hesitate to describe new species on the slender characters that evidentiy were considered sufficient by Hoerold."
Although the original description of C. smithii characterized the leaves as long-acuminate. Smith (1932) emphasized the proportionaUy more oblong, abruptiy caudate-acuminate nature ofthe leaves. Smith said that the leaves of C. smithii averaged 2.2 times longer than broad versus the proportionally narrower leaves of C. costaricensis, C. veraguensis, and C. hoffmannii which averaged 3 times longer than broad. Many of the geographical gaps have now been filled, and the more numerous specimens avaUable and field observations demonstrate great variability in leaf proportion and shape. I do not feel that separate taxa can be maintained on so tenuous a character as leaf proportions without strong additional characters and without regards for geography, elevation, or seemingly habitat.
Smith (1932) further mentioned that C. veraguensis was distinguished from C. costaricensis (and C. hoffmannii) "only by the puberulous calyx, a character which may well be questioned as a basis of specific distinction." When the pubescence patterns of this group of taxa are examined throughout their range, no consistent pattern can be seen. Pubescent plants are found throughout the geographical and elevational ranges, although their frequency increases slightly as one gets above 1780 m in the Central VaUey and above 2400 m in the Cordillera de Talamanca. However, plants observed in the Talamancas growing next to each other have differed only in possessing either pubescent or glabrous twigs. Luteyn 3033 was collected from a pasture at 2800 m elevation on the slopes of Volcan Barba. It consisted of some plants completely glabrous throughout, some with pubescence only on the young twigs, some with pubescence only on the calyces, and others with pubescent calyces and corollas but not twigs. Therefore, within this complex it does not seem that pubescence patterns can be used to differentiate taxa at any hierarchical level.
Cavendishia hoffmannii was distinguished from C. costaricensis by the distinctly punctate leaf surfaces (Smith, 1932). In the material now available it is apparent that these so called "punctate" surfaces are not at all unique to C hoffmannii. In fact they may be seen to a greater or lesser degree on virtually every species of Central American Cavendishia. The punctations occur when the impressed, glandular fimbriae of the leaf surfaces break off and leave a tiny pit or depression in the surface of the leaf. These punctations are most conspicuous on the lower leaf surfaces, and are usually more pronounced in plants taken from higher elevations. However, to distinguish taxa on the basis of these "characters" seems unwarranted.
Cavendishia skutchii was allied with the Mexican-Guatemalan C. crassifolia sensu stricto because of its long rachis and comparatively short corollas. From C. crassifolia it was distinguished primarily by its larger, subcordate leaves which were metallic-colored. In my experience the metallic or slate-blue color found in the leaves of many species of Cavendishia is the result of a combination of the temperature and rapidity by which specimens are dried. It occurs in many taxa of wide distribution. The size and shape ofthe leaves and rachis length have already been discussed as highly questionable characters by which to distinguish taxa. The longer rachis of C. skutchii is rather rare in Costa Rica, although it is not totally absent from the higher elevations of the Cordillera de Talamanca in Costa Rica and Panama. Corolla length is discussed below.
Cavendishia miconioides is a binomial which should be restricted in usage to the South American type although specimens from the vicinity of Volcan Chiriqui were identified as C.miconioides by Smith (1941). Comparisons of the more numerous collections now available from both Chiriqui and Colombia show their distinctive nature (also see the discussion of C. miconioides under doubtful AND EXCLUDED NAMES). The Chiriqui specimens are rather distinctive in their small, oblong or oblong-elliptic leaves. The floral characters are not distinctive, however, and I feel that the leaf characters merely represent one of the extremes in morphological variation, not sufficiently distinct to merit taxonomic recognition.
Cavendishia calycina was characterized by its elongate calyx limb and by its calyx lobes which were oblong, marginally undulate, and basally contiguous. It also possessed short pedicels which were 4-5 mm long, along with correspondingly long, oblong bracteoles about 4 mm long. The taxon has been rarely collected and more collections are needed before an adequate interpretation of this taxon can be made. T h e specimens available show some variation. Although these collections have the calyx characters of C. calycina, the pedicel and bracteole characters mentioned above are restricted to the holotype specimen. Since other collections of C crassifolia from scattered localities throughout the range m a y show tendencies towards elongate calyx limbs and oblong calyx lobes, it seems advisable at this time to synonymize C. calycina with C. crassifolia.
In his key to the Central American taxa. Smith (1932) separated the Mexican-Guatemalan taxa from the Nicaraguan-Costa Rican-Panamanian taxa on the basis of corolla length. Those specimens from the northern part of the geographical range had corollas which supposedly averaged 13-16 mm long and those from the southern part 19 mm long or more (rarely 17 mm long). At that time Smith saw only one collection from the area between Guatemala and Costa Rica. With many more collections available it is apparent that corolla length varies continuously throughout the range, with most corollas being 15-20 mm long. The shortest corollas (12-17 mm long) are most frequently found in Mexico and Guatemala, but m a y also occur in Costa Rica. Longer corollas (18-21 mm long) are more frequently encountered in Costa Rica and Panama, but they are not unknown from the northern part of the range.
With respect to the C. crassifolia complex as a whole, several tendencies may be noted. 1) Rachises greater than 3 cm long are most frequent in the northern part of the range (especially Mexico). Long rachises may also occur at higher elevations in Costa Rica and Panama. Short rachises predominate in the complex as a whole, and especially from Guatemala to Panama, although they may be found in Mexico as well. 2) With respect to coroUa and rachis length considered together, there is seemingly a north-south trend, with short corollaslong rachises being more frequent at the northern limits of the range of the species (Mexico). 3) Plants north of Nicaragua often have bracts which are pubescent along the margins. Very rarely are the bracts of plants from Costa Rica or Panama marginally pubescent. 4) Plants from the northern part of the range (north of Nicaragua) often bear greater numbers of glandular fimbriae upon the calyx and pedicel.
In summary then, the floral characters of the taxa within this complex are virtually identical. Although there are local differences across the geographical range, there are no consistent morphological types associated with realistic geographical ranges, nor are there discontinuities in any combination of characters which could result in a meaningful taxonomy. Therefore, these widely ranging, variable populations are here considered to be members of one polymorphic species, C. crassifolia.
The following collections are suspected to be of hybrid origin between C. endresii and C. crassifolia (see also the chapter on hybridization): COSTA RICA. Cartago: 4.5 km E of bridge at Tapanti above Rio Grande de Orosi, 1400 m, Wilbur & Stone 10499 (DUKE). Heredia: Calle Zurqui, 1500 m , Luteyn 3276 (DUKE, F, MO, NY, WIS), Luteyn et al. 4461 (DUKE). San Jose: vicinity of La Palma, 1290-1350 m, Luteyn 3010 (DUKE, NY), 5570 (DUKE, F, NY).
The following collections are suspected to be of hybrid origin between C. crassifolia and C. pubescens: PANAMA. Chiriqui: Paso Ancho, 1840 m , Luteyn 3053 (DUKE); road between Bambito and Cerro Punta, 1540-1840 m, Luteyn & Wilbur 4652 (DUKE, F, GH, MICH, MO, NY, US), Tyson 5862 (FSU, MO),Wilbur etal. 11879 (DUKE).
-
Distribution
Distribution: Oaxaca, Mexico to Panama; 1000-3200 m elevation. Fig. 37.
Central America|