I-lost Scccc 43(6):1094--1697. 2008. c/ott, varies widely (DArcy and Eshbaugh, 1974: DeWitt and Bosland. 1996: Tewksbury Variation for Fruit Morphological et al., 2006). Recent molecular studies have used C. chinense to elucidate the biochemis- try and the biology of synthesis Characteristics in a cliinense (Stewart et al., 2005). Capsicurn chinense appears to have its Jacq. Germplasm Collection origins in the western Amazon River basin (McLeod et al., 1983) and domesticated Robert Lawrence Jarrett forms appear in early agricultural sites in U.S. Department of Agriculture, Agricultural Research Service, coastal Peru (Davenport, 1970). The oldest Genetic Resources, 1109 Experiment Street, Griffin, GA 30223 known C. c/iinense is a 6500-year-old intact pod found in in Peru Terry Berke (DeWitt and Bosland, 1996). According to SemEn is Vegetable Seeds, 37437 State Highway 16, Woodland, CA 95695 McLeod et al. (1983), the range of C. chi- ne/ice extends throughout central South Additional index words. , fruit morphology, distributions, , evolution America. the , and into Central Abstract. Mature fruit of 330 accessions of (apsicuin chinense America. Heiser (1976) noted the prevalence Jacq. from the USDA/ARS of C. chinense Capsicum germplasm collection were characterized for fruit length, width, weight, and in moister habitats in the color. Mean fruit length was determined to be 47 mm with a range from 7.9 mm to 113.7 lowland tropical areas in . The species mm. Mean fruit width was 21.17 mm with a range of 6.18 mm to 40.0 mm. Mean fruit is also cultivated from and the Caribbean south to and including Peru and weight was 6.31 g with a range of 0.18 g to 22.7g. Distributions of all characteristics were positivel y skewed. Distributions of fruit length, fruit weight, and fruit length/width failed Bolivia (DArcy and Eshhaugh, 1974: Pick- ersgill, 1971), in the Kolmogorov-Smirnov test for normality. The distribution for fruit width was normal. (Bosland et al., Ninety-two percent of the accessions examined were elongate. Mature fruit colors 1996), and in the West Indes (Pickersgill included red, orange, yellow, brown (chocolate), and cream. The gcrmplasm collection et al., 1979). contained primaril y pendent larger-fruited cultivated forms of the species, but also The unique botanical characteristics of C. c/iitiense contained upright small-fruited primitive and semipendent transitional forms. These and the basis for its taxonomic classification are well documented (Smith data define the variabilit y for mature fruit characteristics within this germplasm and Fleiser, 1957). However, other than the collection and provide a baseline against which future introductions/acquisitions can be compared. Additional sampling of the gene pool might be expected to substantially photographs of C. chine//se fruit provided by enhance the genetic variability within the primitive forms of this species. DeWitt and Bosland (1996), the literature contains little if any information on the range of fruit morphological characteristics present in this species. Piekersgill et al. (1979) used Fruit of members of the genus Capsicunt 1996), Smith and Heiser (1957) were the first numerous morphologically distinct forms of play a major nutritional role in many cultures to recognize its ( C sinensis Jacq.) cultivated C. c//jocose in a study to examine phyloge- by serving as a source of vitamin C (Eshbaugh, status. The presence of a calyx constriction is netic relationships within and among the 1976) and other phytonutrients (DeWitt and regarded as the species defining characteris- cultivated taxa. However, descriptions of Bosland, 1996). The pungency associated tic and the one used to differentiate it from individual morphotypes, seed sources, or with many forms of Capsicum makes the C. frutescens. its nearest relative (Eshbaugh, photographs were not provided. The current fresh or dried fruit a desirable spice, and 1976). Capsicwn chinense can be separated study was undertaken to examine and docu- many medicinal properties have been attrib- from C. annteuni on the basis of the number of ment the morphological variation present in uted to capsaicin and its analogs (Stewart pedicels/node: one in C. annuun: and three to fruit of accessions of C. chine,s,ce in the et al., 2005). Although most consumers in the five in C. clunense. Capsicum chinetise is U SDA/ARS Capsicum germplasm collection United States are familiar with the common typically described as having rugose leaves, a (Jarret et al., 1990). sweet (C. annuurn), the pungent dull white or waxy green corolla, and wavy jalapeno (C. an/noun). and Tabasco (Mcllhenny seed margins (Smith and Heiser. 1957). Materials and Methods Co., Avery Island. LA) sauce (C frurescens), the The full potential of C. chinc,i.ce as a food genus Capsicurn actually contains five culti- crop in the United States has not been Seed of a total of 360 accessions, identi- vated species, each with a variety of forms, realized. Eshbaogh (1976) observed that C. fied in the Griffin genebank inventory as few of which are seen frequently in the U.S. chinense rivaled C. annuuni as a crop in parts C. chinen,re, were sown in the greenhouse in marketplace. The genus also contains numer- of South America and the Caribbean. DeWitt May 2005 in Woodland, CA. Of these mate- ous wild species (Hunziker, 2001), several of and Bosland (1996) referred to C. chinense rials, 330 were determined (based on visual which have only recently been described as the most important cultivated pepper in evaluation in the field) to be C. chinetise, the (Barboza and Bianchetti, 2005). South America east of the Andes. DeWitt and remainder (30) being reclassified as either The cultivated species of Capsicum Bosland (1996) also noted the importance of C. aitnutan or C. frute.ccens. The accessions include C. annuu,n. C. baccatu,n, C. frutes- C. chine,i.ce in the Caribbean where fruit of of C. chinense represented germplasm origi- cens. C. pubescens, and C. chinense (Heiser the species was generally referred to as nally acquired from Argentina (one), Belize and Piekersgill, 1969; Smith and Heiser, hahanero, , or goat pepper. (two), Bolivia (22), Brazil (44), Chile (one). 1957). Of these, Capsicum c/unease, often Numerous landraces with specifically adap- China (one). Colombia (24), Costa Rica (17). referred to as hahanero, was the last to be ted fruit types were widely cultivated. In the Cuba (one). Ecuador (21). Guyana (three). recognized as a cultivated taxa in the modern eastern Caribbean, these landraces were Honduras (one), India (two), Mexico (23), scientific literature. Described as early as referred to as Congo peppers (Trinidad) or (one), Peru (II). Philippines (one). 1768 as C. angulosurn (DeWitt and Bosland, honey peppers (Trinidad). In the western (10), Salvador (one), Spain (one). Caribbean, they were referred to as Scotch (nine), Tanzania (two), Trinidad bonnets (Jamaica), rocotillos (Puerto Rico), and Tobago (three), the Untied States (13), Received for publication 22 May 2007. Accepted and cachucha (Cuba). Fruit of all forms of C. (nine), and Zambia (one). Seed- for publication 7 Aug. 2007, chi,iense are generally assumed to be quite lings were transplanted to the field 5 to 6 To whom reprint requests should be addressed; pungent. However, pungency within this weeks after emergence into rows 2 m apart e-mail Bob.Jarret.ars.usda.gov species, like within other species of Capsi- (0.25 m between within rows). Plants

1694 I-IORTSCicNCE VOL. 43(6) OCTOBER 2008 received fertilization, irrigation, weed, and (DArcy and Eshbaugh, 1974: Eshbaugh, distributions for those characteristics were pest control measures, as required. 1976: Smith and Heiser, 1957). Individual not normal. In contrast, values for fruit width Fruit descriptive data were recorded using plants yielded one, two, three, or mixed were normally distributed. Distributions for a modified (in-house) form of the IPGRI numbers of fruits/node. However, a calyx fruit length, weight, and length/width were (1995) descriptor list. Fruit of each genotype constriction was not present on all fruits/plant leptokurtic as opposed to the platykurtic were harvested at full maturity and shipped to or on all plants/accession. Some plants distribution of fruit width (Table 1). Griffin. GA, where they were weighed, mea- exhibited predominantly large solitary flow- DArcy and Eshbaugh (1974) noted that sured, and photographed. Data on fruit ers that were dull white as noted by Pick- fruit of C. chinense were brown (chocolate), length, fruit width, and fruit weight were ersgill et al. (1979). Pis 281442 and 315018 red, tangerine, peach, cherry, neapolitan, recorded on 25 randomly selected sound (and others) yielded fruits that were either red yellow--orange, lemon yellow, or cream. fruit. In those cases in which accessions or chocolate in color but were otherwise near Within the materials examined in the present contained mixtures of two or more fruit types, identical in appearance, suggesting segrega- study, fruit were observed to be various data were taken only on the predominant tion for fruit color (Smith. 1950). In cross- shades of red (143), orange including tarmger- type. Values for each characteristic were section, the internal tissues of some chili me (49), chocolate (20), yellow (16), and averaged within accessions and these means chocolate fruit were red. The fact that this mixtures of red and orange (57), red and further analyzed (Stuart and Ord, 1987). A phenotype was collected from a variety of chocolate (12), red and yellow (II), red list of descriptors. descriptor data, and digital locations over a period of time suggests that and cream (one), or mixtures of three or more images of the fruit of the accessions used in it has been selected for and maintained as colors (26). We observed only a single this study can he viewed at www.ars-grin. a distinct phenotype. Although its taxono- accession with cream-colored fruit, although gov/npgsiaec/ace_queries.html . A spreadsheet mic classification has been questioned the photographs in DeWitt and Bosland containing the data set used in the analysis (Pickersgill et al.. 1979), chili chocolate (1996) suggest that this fruit color is common is available from the authors. types are generally considered to be C. in the gene pool. chinense. Hence, we opted to include these Wild or primitive forms of C. chinen,ce Results and Discussion in the analyses. were also observed. These small, erect- The statistics on the fruit characteristics of fruited weedy forms were described by Although the identification of C. chinense the accessions of C. c/tinense examined are (Pickersgill et al.. 1979) and DeWitt and can be problematic (Davenport, 1970: presented in Table 1. Distributions for the Bosland (1996). The relative lack of recog- Eshbaugh, 1976: Piekersgill, 1966. 1971), characters examined (fruit length, fruit width, nition afforded the primitive forms of C. we feel confident that the materials examined fruit length/fruit width, and fruit weight) are chinense in the modern literature may be were C. chinense based on the clear presence presented in Figure IA- D. Fruit length aver- the result of the paucity of plant material min of a calyx constriction. However, we aged 47.0 ranged from 7.9 mm (P1 available for examination or the suggestion min acknowledge that accessions that have been 446904) to 113.7 Grif 9269). Fruit that C. frute.rcens was the likely wild-type previously referred to in the literature as width averaged 21.2 mm with a range of from progenitor of C. chinense (Pickersgill. 1966, "chili chocolate" (Pickersgill et al., 1979) 6.2 min 260501) to 40.0 mm (Grif 12450). 1971). Pickersgill et al. (1979) recognized are somewhat atypical. The USDA/ARS Average fruit weight was 6.3 g with a range genuinely wild forms of the species with germplasm collection contained 29 acces- of from 0.18 g (111 260504) to 22.7 g (PT small, erect deciduous red fruits and a calyx sions that were classified as (or containing 257136). Fruit length/width averaged 2.4 constriction that were found in the Amazon predominantly) the chili chocolate pheno- with a range from 0.8 (P1 439476) to 6.8 (P1 River basin. Although these primitive forms type. Accessions with a chili chocolate phe- 315031). had been recognized for more than 20 years, notype included P1 nos. 159236, 159246, As indicated in Table 1 and Figure. IA-D, the scientific literature contains no reports of 215736, 257171, 257284, 260532, 281319. the distributions of the values for mature fruit attempts to further characterize them or to 281441, 281442, 315010, 315011, 315012. length, lengthiwidth, and weight were highly further examine their relationship with the 315014, 315015. 315017, 315018, 315021. positively skewed. Assuming that selection cultivated forms of C. chinense or with other 439438. 439440, 439446, 439451, 439441, for fruit size occurred over the course of the wild or cultivated forms of Caps/corn. 439445, 439448, 439449. 639655, and domestication of this species, as seems rela- According to Eshbaugh (1976), all wild 640894 and Griffin Inventory (Grifl nos. tively certain (Eshhaugh. 1980; McLeod peppers have red fruit that are deciduous and 9269 and 9275. The majority of these materi- et al.. 1982: Pickersgill, 1971), positively erect. Indeed, this is typically the case and als were originally acquired by Dr. Paul skewed distributions for these fruit charac- only red-fruited wild-type C. annuurn (C. Smith (formerly of the University of Califor- teristics would be expected because this was annuurn var. glahrioscu!usn) and wild-type nia) and were collected in Chile (one), Peru the organ subjected to selection pressure C.. haccatum (C. haccatuin var. haccafu,n) (22), Spain (one), Trinidad and Tobago (one). (DArcy and Fshbaugh. 1974). Distributions have been observed by RU. In the present the United States (two), and Venezuela (two). of fruit length, fruit length/width, and fruit study. we observed a number of small- of Pickersgill al. (1979) noted that the weight failed the Kolmogorov-Smirnov test fruited, apparently wild-type C. chinense chili chocolate phenotype could not be (Stuart and Ord, 1987) indicating that the with yellow (P1 260501) or orange (P1 referred satisfactorily as either domesticated C. annuurn or domesticated C. c/linense. We observed that accessions of chili chocolate Table I. General statistics on four fruit parameters among 330 accessions of C (IpSi(iiflt (hinemc, had a plant growth habit and general appear- Fruit length Fruit width Fruit Fruit ance that resembled C. annmoim in some Statistic (mlii) (mm) wt ( gi tengthiss idtti respects and that plants and fruits were quite Mean 47.035 21.171 6.311 2.379 uniform across accessions regardless of on- SE) 19.196 6.839 4.131 1.088 gin. Stems and leaves of many, but not all, sr. 1.057 0.376 0.227 0.059 chili chocolate accessions were noticeably Cl of mean 2.079 0.741 0.447 0.118 pubescent. Leaves were typically not rugose. Range (05.80 33188 22.535 6.019 Maximum 113.70 40.000 22.710 6.830 Mean (±sD) fruit length, width, and weight Minimum 7.900 6.182 0.175 0.811 across all 29 chili chocolate phenot ypes were Median 44.00 21.275 5.580 2.129 61.5 mm (11.1), 22.9 mm (3.2). and 10.61 g Skewness 0.785 0.175 0.832 1.1 It) (0.94), respectively. In all eases, fruit were Kurtosis 0.620 -0.513 0.459 1.404 firm at maturity, elongate, and pendent and K-S distribution 0.088 0.0408 0.095 0.1t16 corollas were usually greenish white, char- Failed. acteristics associated with C. chinense Cl. confidence interval; K-S. Kolmogorov-Smirnov.

HoioScasci: VE)t . 43(6) OcTonra 2008 1695 25 16

14

20 12

U) U) 10 .9 15 U) U) U) aU) C-) C-) C-) a CU d 10 06 z Z 4 5 2

0 0 0 20 40 60 80 100 120 0 10 20 30 40 50 Fruit length (mm) Fruit width (mm) 25 30

25 20

U) U) 20 C .2 15 0 U) U) U) U)

C-) 15 C-) 0 CU CU 10 0 Z Z 10

5 5

0 0 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 Fruit weight (g) Fruit length/width

Fig. 1. Distribution of fruit length (A), width (B), weight (C), and length/width (D) values of 330 accessions of C. c/unenxe.

260504) fruit that were small, deciduous, and an intergrading and poorly differentiated polymorphisms to examine genetic relation- erect. Although wild-type fruit are typically complex, which could not be readily divided ships among four Capsicum sp., including red, it has also been acknowledged that into distinct taxa, whereas the domesticated 89 accessions of C. chinense. However, the variation for morphological characteristics forms of each taxon were clearly distinguish- morphological attributes of those accessions within wild taxa may range from a typical able from one another. This was interpreted were not noted. Buso et al. (2003) used wild type to a semi- or fully domesticated as being suggestive of an independent allo- RAPDs to examine genetic differences form (Eshbaugh, 1976); this would include patric domestication of the four species within and among various Capsicum sp. fruit color. followed by divergence of the after Although II accessions of C. chinense were Small-fruited accessions such as P1 domestication. included, data on the primitive versus culti- 260504 were relatively rare in the collection Few molecular studies have addressed the vated status of the materials analyzed was not when compared with the larger pendent relationship between C. chinense phenotypes provided. forms. This may be attributable to the more with fruit that are small, persistent, and Many of the characteristics that are used defined (restricted) geographic range of these upright and their pendent large-fruited rela- to define "wild-type" C. chinense (small, types combined with a less intense effort to tives. Prince et al. (1995) used random upright, small, red, and nonpersistent fruit) collect and conserve them. The distribution amplified polymorphic DNAs (RAPDs) and are essentially those same characteristics of these primitive forms appears to be largely restriction fragment length polymorphisms to (i.e., fruit size, position, size, color, and undocumented. To the authors knowledge, examine DNA polymorphisms within Capsi- shape) that have been previously acknowl- systematic studies that included both primi- cum but did not address the issue of wild edged as being unreliable for species delim- tive and cultivated forms of C. chinense have versus domesticated forms. RAPD markers itation (Pickersgill et al., 1979). However, not been conducted with the exception of useful in identifying individual Capsicum this does not preclude their use in describing Pickersgill et al. (1979). The electrophoretic species were reported by Rodriguez et al. varieties. Pickersgill (1966) and Eshbaugh studies of Jensen et al. (1979), McLeod et al. (1999). However, these were not applied to (1980) recognized that the parallel develop- (1979), and McLeod et al. (1983) included C. both the wild and cultivated forms. Walsh ment of C. annuu,n. C. Ji-utescens, C. bacca- chinense. However, no distinction was made and Hoot (2001) examined phylogenetic turn, and C. chinense would lead to erroneous between cultivated and wild forms. The study relationships in Capsicum using DNA conclusions regarding subspeciflc classifica- of Pickersgill et al. (1979) included a single sequence polymorphisms in the chloroplast tion using morphological characteristics. wild-type C. chinense and a number of atpB-rhcL spacer region and the (nuclear) DArcy and Eshbaugh (1974) suggested that transitional forms. They (Pickersgill et at., waxy intron. However, only a single acces- different names for the cultivated and wild 1979) noted that the wild forms of C. ann- sion of C. chinense was included. Toquica forms, at some taxonomic level, were clearly uum, C. chinense, and C. frutescens formed et al. (2003) used amplified fragment length required. Eshbaugh (1980) suggested that

1696 HORTSCIENCE VOL. 43(6) OCTOBER 2008 the taxonomic logic used in distinguishing Heiser, C.B., Jr. 1976. Pepper-Capsicwn (Solana- A.D. Skelding (eds.). The biology and taxon- C. annuum var. annuum from C. annum var. ceae), p. 265-268. In: Simmonds, N.W. (ed.). omy of the . Linn Soc. Symp. Ser. No. 7. Academic Press, London, UK. glabriusculum also be applied to designate Evolution of crop plants. Longman. London, UK. Heiser, C.B.. Jr. and B. Pickersgill. 1969. Names Prince. J.P., V.K. Lackney, C. Angeles, J.R. the wild ancestral and cultivated forms of for the cultivated Capsicum species (Solana- Blauth, and M.M. Kyle. 1995. A survey of C. chinense. This has yet to be done. The use ceae). Taxon 18:277-283. DNA polymorphism within the genus Cap.ci- of varietal epithets does provide a means to Hunziker, AT. 2001. Genera Solanacearum: The corn and the fingerprinting of pepper cultivars. distinguish the morphological extremes of genera of the Solanaceae illustrated, arranged Genome 38:224-231. the species in terms of its fruit characteristics. according to a new system. A. R. G. Gantner Rodriguez, J.M., T. Berke, L. Engle. and J. Nuihuis. However, the line between wild and semi- Verlag K.-G., Ruggell. Liechtenstein, Ger- 1999. Variation among and within Capsicum domesticated forms and between semidomes- many. species revealed by RAPD markers. Theor. ticated and cultivated forms can be obscure. IPGRI. 1995. Descriptors for Capsicum (Capsicum Appl. Genet. 99:147-156. Smith, P.G. 1950. Inheritance of brown and green This can confound efforts to provide an spp.). International Plant Genetic Resources Institute, Rome, Italy. fruit mature fruit color in peppers. J. Hered. accurate subspecific classification. Jarret, R.L., M. Spinks, G. Lovell, and A.G. 41:138-140. Smith, P.G. and C.B. Heiser, Jr. 1957. Tax- Literature Cited Gillaspie. 1990. The S-9 plant germplasm collections at Griffin. GA. Diversity 6:23-25. onomy of Caps/corn sinense Jacq. and the Barboza, G.E. and L.De.Bern Bianchetti. 2005. Jensen, R.J., M.J. McLeod. W.H. Eshhaugh, and geographic distribution of the cultivated Cap- Three new species of Capsicum (Solanaceae) S.I. Guttman. 1979. Numerical taxonomic sicurn species. Bull. Torrey Bot. Club 84:413- and a key to the wild species from Brazil. Syst. analyses of allozymic variation in Capsicum 420. But. 30:863-871. (Solanaceae). Taxon 28:315-327. Stewart, C., Jr., B.-C. Kang, K. Liu, M. Mazourek, Bosland, P.W., A.T. Bailey, and J. Iglesias-Olivas. McLeod, Mi., W.H. Eshbaugh, and S.I. Guttmam. S.L. Moore, E.Y. Yoo, B.-D. Km, 1. Paran. and 1996. Capsicum pepper varieties and classifi- 1979. A preliminary biochemical study of the M.M. Jahn. 2005. The pun l gene for pungency cation. New Mexico State University, Cooper- genus Capsicum --Sol anaceae, p. 701-713. In: in pepper encodes a putative acyltransferase. ative Extension Service, Circular 530. Hawkes, J.G., R.N. Lester. and A.D. Skelding Plant J. 42:675-688. Buso, G.S.C., Z.P. de Sousa Amaral, L. Dc. Bern (eds.). The biology and taxonomy of the Sol- Stuart. A. and J.K. Ord. 1987. Kendalls advanced Bianchetti, F.B.B. Machado, and M.E. Ferreira. anaceae. Academic Press, New York, NY. theory of statistics. Charles Griffin and Co., 2003. Genetic variability and phylogenetic McLeod, Mi., S.I. Guttman. and W.H. Eshbaugh. London. UK. analysis of Brazilian species of Capsicum. 1982. Early evolution of chili peppers (Capsi- Tewksbury, ii., D.J. Levy, D.C. Haak, and C. Capsicum Eggplant Newslett. 22:13-16. cum). Eco. Bot. 36:361-368. Manchego. 2006. Where did the chili gets its DArcy. W.G. and W.H. Eshbaugh. 1974. McLeod, Mi., S.I. Guttman, W.H. Eshbaugh. and spice? Biogeography of capsaicinoid produc- New World peppers [Capsicurn-Solanaceae] R. Rayle. 1983. An electrophoretic study of tion in ancestral wild chili species. J. Chem. north of Colombia: A resume. Baileya 19:93- evolution in Capsicum (Solanaceae). Evolution Ecol. 32:547-564. 105. Int. J. Org. Evolution 37:562-574. Toquica, S.P., F. Rodriguez, E. Martinez, M.C. Davenport, W.A. 1970. Progress report on the Pickersgill, B. 1966. The variability and relation- Duque, and J. Tohme. 2003. Molecular char- domestication of Capsicum (chili peppers). ships of Capsicum chinense Jacq. Indiana acterization by AFLP5 of Capsicum germplasrn DeWitt, D. and P.W. Bosland. 1996. Peppers of the University, Bloomington, IN. PhD Diss. from the Amazon Department in Colombia, world. Ten Speed Press., Berkeley, CA. Pickersgill, B. 1971. Relationships between weedy characterization by AFLPs of Capsicum. Eshbaugh, H.W. 1980. The taxonomy of the genus and cultivated forms in some species of chili Genet. Res. Crop Evol. 50:639--647. Capsicum (Solanaceae). Phytologia 47:153- peppers (genus Capsicum). Evolution Int. J. Walsh, B.M. and S.B. Hoot. 2001. Phylogenetic 166. Org. Evolution 25:683-691. relationships of Capsicum (Solanaceae) using Eshbaugh, W.H. 1976. XII. Genetic and biochem- Pickersgill, B., C.B. Heiser, and J. McNeill. 1979. DNA sequences from two noncoding regions: ical systematic studies of chili peppers (Capsi- Numerical taxonomic studies on variation and The choloroplast ATPB-RBCL spacer region cuni-Solanaceae). Bull. Torey Bot. Club. domestication in some species of Capsicum, p. and nuclear wan introns. Int. J. Plant Sci. 102:396-403. 679-700. In: Hawkes, J.G., R.N. Lester. and 162:1409-1418.

HORTSCIENCE VOL. 43(6) OCTOBER 2008 1061TA