| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics |
International Potato Center (CIP), Apartado 1558, Lima 12, Peru; United States Department of Agriculture, Agricultural Research Service, Department of Horticulture,University of Wisconsin, 1575 Linden Drive, Madison, Wisconsin 53706-1590 USA
Received for publication July 26, 2001. Accepted for publication January 3, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: cultivated • nomenclature • phenetic • potato • sect. Petota • Solanaceae • Solanum tuberosum • taxonomy
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Table 1). The cultivated potato taxonomy of the International Potato Center (CIP) and the United States potato genebank (NRSP-6) follows Hawkes (1990)
because it is the latest comprehensive treatment and because he identified many of the cultivated species at CIP and NRSP-6. In addition to the cultivated species there are 199 tuber-bearing wild species relatives, distributed from the southwestern United States to south-central Chile (Spooner and Hijmans, 2001
).
|
treatment of the seven cultivated species is not universally accepted and is part of a long history of disagreement among potato taxonomists of the treatment of both cultivated and wild species (Spooner and van den Berg, 1992
). For example, the Russian potato taxonomists Bukasov (1971)
and Lechnovich (1971)
recognized 21 species, including separate species status for S. tuberosum subsp. andigenum (as S. andigenum) and subsp. tuberosum (as S. tuberosum). Ochoa (1990
, 1999
) recognized nine species (Table 2) and 141 infraspecific taxa (subspecies, varieties, and forms; including his unlisted autonyms) for the Bolivian cultivated species alone.
|
proposed that modern cultivars of Solanum tuberosum subsp. tuberosum originated from landraces from Chile. Hawkes (1990)
and Hawkes and Francisco-Ortega (1993)
suggested an origin of modern cultivars from Andean landraces of S. tuberosum subsp. andigenum, with later breeding with Andean landraces, Chilean landraces, and wild species. Hawkes (1990)
classified Chilean landraces and modern potato cultivars under Solanum tuberosum subsp. tuberosum, as they have attained a similar morphology of wide leaflets and adaptation to flowering and tuberizing under long-day conditions.
Bukasov (1971)
, Lechnovich (1971)
, Hawkes (1990)
, and Ochoa (1990)
classified potatoes as distinct species under the International Code of Botanical Nomenclature (ICBN; Greuter et al., 1999
). Dodds (1962)
, in contrast, treated the cultivated species under the International Code of Nomenclature of Cultivated Plants (ICNCP; the latest version is Trehane et al., 1995
). He suggested that there was poor morphological support for most cultivated species, and recognized only S. xcurtilobum, S. xjuzepczukii, and S. tuberosum, with five "groups" recognized in the latter (Table 2). "Cultivar-groups" (the current terminology) are taxonomic categories used by the ICNCP to associate cultivated plants with traits that are of use to agriculturists. The cultivar-group classification of Dodds (1962)
was based on comparative morphology, reproductive biology, cytological and genetic data, and cultural practices. He contended that the morphological characters used byHawkes (1956a, b)
to separate species exaggerated the consistency of qualitative and quantitative characters. He showed that Andean farmers grow landraces of all ploidy levels together in the same field and that these can all potentially hybridize. He showed no genetic differentiation of the cultivated diploids (Dodds and Paxman, 1962
). He contended that his classification was conservative in that it "provides a genetically reasonable classification that disturbs the established usage of words [taxonomic names] as little as possible" (Dodds, 1962
, p. 530).
Later data supported Dodds' (1962)
hypothesis of poor morphological separation of the cultivated species and suggested that they formed a genetically diverse assemblage of genotypes of multiple and complex hybrid origins. Despite the contention of Jackson, Hawkes, and Rowe (1977)
that there was limited gene flow between diploid and tetraploid cultivated species, Hawkes (1990)
proposed that the triploid S. chaucha was of hybrid origin between the diploid species S. phureja subsp. phureja or S. stenotomum and the tetraploid species S. tuberosum subsp. andigenum. Many studies have shown that potato fields in the Andes contain mixtures of cultivated species at different ploidy levels (Ochoa, 1958
; Huamán, 1975
; Jackson, Hawkes, and Rowe, 1980
; Brush, Carney, and Huamán, 1981
; Johns and Keen, 1986
; Johns et al., 1987
; Quiros et al., 1990
, 1992
; Zimmerer, 1991
). Cultivated species frequently co-occur with different wild potato species (Ugent, 1970
; Huamán, 1975
; Grun, 1990
). The boundary between "cultivated" and "wild" is often vague, and some putative "wild" species may be revertants from cultivation (Spooner et al., 1999
). Watanabe and Peloquin (1989
, 1991
) showed both diploid and unreduced gametes to be common in the South American wild and cultivated species, potentially allowing gene transfer among different ploidy levels. Huamán (1975)
showed evidence of natural crosses between the diploid wild species S. megistacrolobum and the diploid cultivated species S. stenotomum. Open pollinated hybrid fruits were found in all experimental plots containing 10, 25, 50, and 90% of S. megistacrolobum plants within isolated plots of S. stenotomum grown in Huancayo, Peru. Rabinowitz et al. (1990)
tested hypotheses of gene flow between the diploid wild taxon S. sparsipilum subsp. sparsipilum and the diploid cultivated species S. stenotomum. By use of isozyme markers specific to these populations, they were able to document high levels of natural gene flow in experimental field plots in the Andes. Tay (1979)
showed extensive overlap of ranges of character states between S. stenotomum subsp. stenotomum and subsp. goniocalyx and questioned their treatment as distinct taxa. Hawkes and Hjerting (1989
, p. 376) questioned the distinctness of all three diploid taxa (S. phureja, S. stenotomum subsp. stenotomum, S. stenotomum subsp. goniocalyx), except S. ajanhuiri. They suggested that "in the future they well may need to be classified entirely under S. stenotomum, with a subspecies distinction for S. goniocalyx, and perhaps also for S. phureja" (Hawkes and Hjerting, 1989
, p. 376). Hawkes and Hjerting (1989
, p. 388) recognized S. chaucha (triploid) despite their statement that "it is merely a convenient label for a series of nothomorphic forms resulting from many crosses between various clones of its parental species."
The ICNCP groups cultivated plant names under denomination classes. A denomination class is a nomenclatural device found in the ICNCP, not the ICBN. It is defined (ICNCP Arts. 6.1, 17.2) as a taxon, or a designated subdivision of a taxon, or a particular cultivar-group, within which cultivated plant epithets must be unique. The botanical genus is the denomination class by default, but Solanum tuberosum is the accepted denomination class for cultivated potatoes (Trehane et al., 1995
, p. 68). A cultivar epithet must only exist once in every denomination class (Spooner et al., 2002
).
Classification and nomenclature of cultivated plants can follow rules of the ICNCP or the ICBN, and classifications of cultivated potatoes in one or the other may reflect differing hypotheses about their evolutionary dynamics. One hypothesis, presented by Ugent (1970
; Fig. 1a) postulated extensive gene flow within and among ploidy levels of cultivated and wild species (the crop weed concept; e.g., Harlan, 1992
), precluding maintenance of species. Conversely, Hawkes (1962
, 1990
; Fig. 1b) postulated gene flow to lead to stabilized hybrids and for further hybrids to be eliminated by reduction in fertility in advanced hybrid generations.
|
was the first to count chromosomes of the cultivated potatoes and discovered diploids, triploids, tetraploids, and pentaploids and used these data to speculate on their hybrid origins. In historical and current practice, identifications are frequently made only after chromosome counts are determined, and reidentifications made after chromosome counts do not match that expected for the species. The strong reliance on ploidy levels was clearly stated by Hawkes and Hjerting (1989
, p. 389): "The chromosome number of 2n = 36 largely helps to identify S. chaucha, but morphological characters can also be used."
As taxonomists responsible for the identification and nomenclature of major germplasm holdings of cultivated and wild potatoes, we need to resolve these disagreements in taxonomy. For the wild potato species, our goal is to produce a more stable and natural classification and to pursue monophyletic taxa (e.g., Baum and Donoghue, 1995
). Our impression of extensive morphological intergradation among the cultivated species and knowledge of literature (above) led us to suspect, however, that the cultivated species were not monophyletic. The ICNCP recognizes the complex hybrid origins of most crops and focuses on a classification of convenience to users and nomenclatural stability needed for trade (Hetterscheid and Brandenburg, 1995
). These are important practical goals. Potato genetic resources provide resistances, sometimes of an extreme type, to the pests and diseases affecting cultivated potato and are sources of improved agronomic traits (Ross, 1986
; Hawkes, 1990
; Spooner and Bamberg, 1994
; Huamán, Golmirzaie, and Amoros, 1997
). Publications reporting use of potato genetic resources appear monthly in the scientific literature. The treatments of Dodds (1962)
andHawkes (1990)
continue to be widely used as parallel but competing taxonomic systems and maintain confusion among users and instability in taxonomy.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). We chose up to 48 morphologically most distinct collections from each taxon as recognized by Hawkes (1990)
(Tables 1 and 3), if this number was available from the collection (S. ajanhuiri had 7 accessions, S. chaucha had 37, S. curtilobum had 4, S. juzepczukii had 14, S. phureja subsp. phureja had 43, S. stenotomum subsp. goniocalyx had 32, S. stenotomum subsp. stenotomum had 42, S. tuberosum subsp. andigenum had 48, S. tuberosum subsp. tuberosum had 30, and putative hybrids between S. stenotomum subsp. S. goniocalyx and S. stenotomum subsp. stenotomum had 10). These include all species and subspecies of Hawkes (1990)
, except the rare and localized taxa S. phureja subsp. estradae and subsp. hygrothermicum (Table 1) that were not available for analysis. All accessions of S. tuberosum subsp. tuberosum are landrace populations from Chile, not modern commercial cultivars.
|
; Huamán et al., 2000
). As outlined in Huamán, Golmirzaie, and Amoros (1997)
, CIP had only 10 accessions for S. ajanhuiri, 31 for S. juzepczukii, and 11 for S. curtilobum. However, the number of distinct genotypes of S. ajanhuiri in this collection was thought to be only seven (Huamán, Hawkes, and Rowe, 1980
), for S. juzepczukii 21 (Schmiediche, Hawkes, and Ochoa, 1980
), and for S. curtilobum 2 (Schmiediche, Hawkes, and Ochoa, 1980
). Huamán (unpublished data) has observed five morphologically distinct types of the latter based on tuber skin color and sprout color, and we include four here. Five tubers of each accession were planted directly in a field plot in Huancayo, Peru (elevation 3200 m above sea level [a.s.l.], 12° S, 75° W). Vouchers were deposited at the herbarium of the International Potato Center in Lima, Peru (herbarium code CIP).
|
). The middle of the five plants per row was measured per accession. We assessed 74 morphological characters; 54 were quantitative and 20 were qualitative (Table 3). We also assessed one developmental character used to define S. phureja (tuber dormancy at harvest, character 75, Table 3). Of the 54 quantitative characters, 17 were ratios to assess shapes, and none of these ratios weighted characters by using a character more than once. The 74 morphological characters assessed all 13 morphological characters mentioned in keys of Hawkes and Hjerting (1989)
, Hawkes (1990)
, and Ochoa (1990)
to distinguish the species (Table 4). The raw data file is deposited at http://ajbsupp.botany.org/v89/.
|
used the angle of divergence of the leaf from the stem to distinguish S. tuberosum subsp. andigenum from S. tuberosum subsp. tuberosum. To thoroughly assess this trait we measured this character on two different portions of the stem (characters 36 and 37). Floral characters were measured on the uppermost inflorescence. Corolla and tuber colors were measured using color charts that were designed at CIP (Huamán and Gómez, in press) and are available from the authors. Corolla colors are arranged from white to pink, to blue to lilac to violet, and intensities of these colors. The CIP color codes (in parentheses) and equivalent color codes from the Royal Horticultural Society (1986)
[in brackets], are: white [155D], cream [10C], light yellow [5A] (1.1, 1.2, 1.3); light pink [65D], medium pink [68D], dark pink [57C] (2.1, 2.2, 2.3); light red [61C], medium red [67A], red-purple [71B] (3.1, 3.2, 3.3); light blue [108A], medium mauve [100D], dark mauve [102D] (4.1, 4.2, 4.3); medium blue [105B], dark blue [105A], blue-purple [94A] (5.1, 5.2, 5.3); light lilac [76C], medium lilac [84B], dark lilac [86D] (6.1, 6.2, 6.3); medium red purple [72A], darker red-purple [77A], dark red-purple [81A] (7.1, 7.2, 7.3); medium-purple [83B], dark purple [86A], dark violet [89A] (8.1, 8.2, 8.3). Tuber colors are explained in Table 3. Ratio characters 12 and 13 are different measures of leaf shape, and we used character 13 to test ideas of Salaman (1949)
regarding morphological differences of the subspecies of S. tuberosum (long narrow leaves characteristic of subsp. andigenum, broad condensed leaves characteristic of subsp. tuberosum).
Data analysis
Quantitative characters were analyzed for their means, ranges, and standard deviations. Character distributions among taxa were determined in JMP statistical software (SAS, 1995
) by the Tukey-Kramer honestly significant difference (HSD) test. Dendrograms including all accessions were produced by NTSYS-pc version 1.70 (Rohlf, 1992). Means for each character were standardized (STAND) and similarity matrices (in SIMINT) were generated using product-moment correlation (CORR), average taxonomic distance (DIST), Euclidean distance (EUCLID), and Manhattan distance (MANHAT). Clustering was performed using the unweighted pair-group method (UPGMA). Cophenetic correlation coefficients (COPH and MXCOMP) were used to measure distortion between the similarity matrices and the resultant three phenograms (Rohlf and Sokal, 1981; Sokal, 1986
). Principal components analysis (PCA) was run with NTSYS-pc, and SAS, version 7 (1998). The PCA with NTSYS-pc used STAND, CORR, and EIGEN. Stepwise discriminate analysis (SDA) was performed with SAS using STEPDISC. Canonical discriminate analysis (CDA) was performed with SAS using CANDISC.
The PCA was performed three times, once with all 267 taxa and 75 characters, with NTSYS-pc that handles missing data. Because of the potential effect of missing data on phenetic results, we ran two additional analyses with no missing data cells, each constructed by elimination of characters or taxa. None of the accessions of S. tuberosum subsp. tuberosum produced flowers or fruits in Huancayo, Peru, so one analysis was run with all accessions but with the 25 floral characters and two fruit characters (Table 3) deleted for all taxa. Another analysis was made with 201 accessions that excluded the two fruit characters that were the most common missing data across all accessions and eliminated all 30 accessions of S. tuberosum subsp. tuberosum and 36 accessions of other taxa (also lacking flowers and fruits). Stepwise discriminate analysis was run with SAS on this 201 accession data set.
The SAS does not analyze any accessions with any missing data, and the CDA also was performed twice, using the same reduced data sets as above focused on the elimination characters or taxa. In addition, both these reduced dataset CDA analyses needed to be conducted with the elimination of two characters with only two character states that were invariant within taxa (pedicel articulation, character 45; tuber dormancy at harvest, character 75).
The PCA and CDA are both ordination techniques, but PCA makes no assumptions about group membership of OTUs. It attempts to portray multidimensional variation in the data set in the fewest possible dimensions, while maximizing the variation. The CDA uses assigned groups to derive a linear combination of the variables (morphological characters) that produces the greatest separation of the groups (SAS, 1998
) and is a much more powerful technique than PCA to separate groups. Cluster analysis, like PCA, makes no assumptions about group membership; it produces trees based on average similarity of all data. The PCA and dendrograms, therefore, are more appropriate to explore phenetic structure without any assumptions of species boundaries, while CDA is an appropriate technique to test preexisting classifications.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
The Tukey-Kramer HSD test determined that all the characters were significantly different (P = 0.05) between at least two taxa. We show the means, ranges, and standard deviations of 42 of these 75 characters in Fig. 3. We chose them based on using 26 characters we consider to best assess components of characters used in past treatments (listed with an asterisk on Table 3), and all 38 characters of the SDA supported as distinguishing taxa (providing 16 additional characters considering duplicates of these two classes of characters). All characters are highly polymorphic, and the only absolute species-specific character state is for S. phureja subsp. phureja, distinguished by the one physiological character used in this study (tubers sprouted at harvest).
|
|
A PCA of one of the two reduced data sets (all accessions, only 51 of the 75 characters, no floral or fruit characters, no missing data cells) is presented in Fig. 5. This PCA most clearly separates S. tuberosum subsp. tuberosum, as in Fig. 4, showing that its morphological distinction is not an artifact of missing floral and fruit characters and that it is supported by vegetative characters. This PCA also provides some morphological support for most accessions of S. juzepczukii, but S. chaucha and S. phureja subsp. phureja are less well supported than in Fig. 4.
|
|
A CDA of one of the two reduced data sets (all accessions, only 49 of the 75 characters) is presented in Fig. 7. This CDA continues to separate S. tuberosum subsp. tuberosum and this time not intermixed with a couple of accessions of S. stenotomum subsp. stenotomum, as in Figs. 4 and 5. It provides better morphological support for S. ajanhuiri and S. juzepczukii than in Fig. 4.
|
|
and Ochoa (1990)
overlap in range so much with those of other taxa that they have no or greatly reduced use in key construction. For example, both authors distinguished S. phureja subsp. phureja from S. stenotomum (both subspecies) by the fact that subsp. phureja has shiny leaves and S. stenotomum has dull leaves. Our results show S. phureja subsp. phureja to be shiny only 65% of the time and S. stenotomum subsp. stenotomum to be shiny 40% of the time. Hawkes (1990)
distinguished S. stenotomum subsp. goniocalyx by tubers with bright yellow flesh, and we found only 16% with bright yellow flesh, 34% with yellow flesh, and 19% with pale yellow flesh. However, S. phureja subsp. phureja showed similar frequencies of yellow flesh to those of S. stenotomum subsp. goniocalyx. Hawkes (1990)
also distinguished S. stenotomum subsp. goniocalyx from S. stenotomum subsp. stenotomum by the fact that subsp. goniocalyx has a ribbed calyx base and subsp. stenotomum has an unribbed base. We found S. stenotomum subsp. goniocalyx with ribbed bases 88% of the time, while S. stenotomum subsp. stenotomum had them 10% of the time. A ribbed calyx is also present in S. phureja subsp. phureja 79% of the time. Hawkes (1990)
distinguished S. tuberosum subsp. tuberosum from S. chaucha, S. phureja subsp. phureja, and S. stenotomum by suggesting that S. tuberosum subsp. tuberosum had symmetrical calyces and S. chaucha, S. phureja subsp. phureja, and S. stenotomum had irregular calyx arrangements arranged as 2 + 2 + 1 or 2 + 3 groups. Although S. tuberosum subsp. tuberosum did not flower in our study, S. chaucha was regular 87% of the time; S. phureja subsp. phureja was 37%, S. stenotomum subsp. stenotomum was 29%, and S. stenotomum subsp. goniocalyx was regular 41% of the time. Hawkes (1990)
recognized S. ajanhuiri, S. curtilobum, and S. juzepczukii as having the terminal leaflet not arched at the tip and all other taxa as having the terminal leaflet arched. Our data showed S. ajanhuiri and S. curtilobum to have the terminal leaflet to be slightly arched downward 100% of the time. Solanum juzepczukii showed 77% of leaflets slightly arched downward and 23% straight tips. All other species had from 66 to 89% arched downward and 11–34% straight leaf tips. Similarly, Hawkes (1990)
considered S. ajanhuiri to be distinguished by its pentagonal corollas. Our measure of corolla dissection (ratio of corolla lobe width to lobe length; character 53 in Table 3) ranged from 1.5 to 2.0 in S. ajanhuiri, and similar ratios were found in both subspecies of S. stenotomum (60%) and in S. tuberosum subsp. andigenum (55%). The same ratio for other species range between 2.25 and 3.25 that include rotate and very rotate corollas. Hawkes (1990)
distinguished S. juzepczukii by its smaller flowers. However, the corolla diameter of S. juzepczukii ranged from 3.0 to 4.0 cm, and we found similar sizes in S. tuberosum subsp. andigenum 21% of the time, S. phureja subsp. phureja 56%, S. stenotomum subsp. stenotomum 53%, and S. stenotomum subsp. goniocalyx 75% of the time. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
) and ser. Longipedicellata (Spooner, van den Berg, and Miller, 2001
).
For the wild species, our goal is to recognize monophyletic taxa (e.g., Baum and Donoghue, 1995
), with a realization of the likely need to recognize some plesiospecies (e.g., Rieseberg and Brouillet, 1994
; Olmstead, 1995
). Our data and other data presented above, however, suggest that the cultivated species are of complex hybrid origins, often intergrade morphologically, and are better classified under the rules of the ICNCP that recognizes these phenomena as typical of crops and focuses on a classification of stability and convenience to users (Hetterscheid and Brandenburg, 1995
; Spooner et al., 2002
). While S. ajanhuiri, S. chaucha, S. curtilobum, S. juzepczukii, and S. tuberosum subsp. tuberosum show some degree of morphological support, we consider treatment as cultivar-groups by the ICNCP to be a more appropriate classification. These "taxa" have common progenitors and/or common hybrid origins (Fig. 1) and undergo hybridization with wild and weedy taxa (except landrace populations of S. tuberosum subsp. tuberosum, which are geographically isolated from other wild potatoes). Many of the cultivated species likely were selected many times from members of the wild species in the S. brevicaule complex (Ugent, 1970
; Grun, 1990
; Hosaka, 1995
; van den Berg et al., 1998
; Miller and Spooner, 1999
). The distinction between wild and cultivated species is often vague, and some putative wild species could equally be progenitors or escaped cultivated species. Indeed, some accessions of both groups are so similar that classification as cultivated or wild often rests on whether they are collected in the wild or in a cultivated field (Spooner et al., 1999
).
We agree with Hetterscheid and Brandenburg (1995)
, who argue that the ICBN should be reserved exclusively to name wild species, for which there is a better chance of discerning evolutionary relationships. This is more problematical for cultivated species because of more natural and artificial hybridization, movement of germplasm away from its natural geographic ranges and habitats, and rapid morphological change through artificial selection. Nomenclature in the ICBN is portrayed as a series of nested classification ranks. Each higher rank (form, variety, subspecies, species, genus, etc.) contains the members of lower ranks, and membership in these ranks implies phylogenetic relationships. These authors advocate classification of cultivated plants as "culta," not taxa, where no attempt is made to group cultivated plants in classifications implying phylogenetic relationships, except within larger taxa that are part of the denomination class (here Solanum, as Solanum tuberosum is the denomination class). Our proposed classification places all cultivated populations as cultivar-groups of the single denomination class S. tuberosum.
Because of some phenetic support (Figs. 4–8), a reasonable argument can be made to recognize S. ajanhuiri, S. chaucha, S. curtilobum, S. juzepczukii, and S. tuberosum subsp. tuberosum as separate species or subspecies, and all the other taxa as cultivar-groups under a separate cultivated species Solanum andigenum. Support for a separate taxon treatment is provided by Raker and Spooner (2002)
who demonstrate that most of the landrace populations of S. tuberosum subsp. tuberosum can be distinguished with microsatellite data from most populations of S. tuberosum subsp. andigenum, and we expect that molecular support will be provided for S. ajanhuiri, S. curtilobum, and S. juzepczukii. Distinct species status also could be argued for S. ajanhuiri, S. curtilobum, and S. juzepczukii by their separate hybrid origins involving the phenetically distinct wild species S. acaule Bitter or S. megistacrolobum Bitter (Fig. 1b). Solanum phureja, S. stenotomum (both subspecies), and S. tuberosum subsp. andigenum, on the other hand, possibly evolved from members of the S. brevicaule complex (S. leptophyes, S. sparsipilum) and the distinction between these wild and cultivated species is often vague (van den Berg et al., 1998
; Miller and Spooner, 1999
; Spooner et al., 1999
). Separate species status also could be reasonably argued by a classification philosophy that focuses on a phenetic rather than a cladistic criteria to define taxa, as argued by McNeill (1998)
and in review of our paper.
We classify cultivated species under the single denomination class S. tuberosum because of their predominant polythetic morphological support, reticulate origins (Hawkes, 1990
; Huamán, Hawkes, and Rowe, 1980
, 1982
, 1983
; Schmiediche, Hawkes, and Ochoa, 1982
; Cribb and Hawkes, 1986
), possible multiple origins involving common species (Hosaka, 1995
), evolutionary dynamics of continuing hybridization, and our classification philosophy of the appropriateness of the ICNCP for cultivated species.
Our proposed classification does not provide synonymy of the many species names to these cultivar-groups. This is an unfinished nomenclatural task because many names published in the Russian literature are of dubious nomenclatural standing and have yet to be typified. However, Hawkes (1990)
and Ochoa (1990)
list many cultivated species synonyms. The association of species epithets to our cultivar-group names is clear by their similarity of names. We consider most accessions of S. stenotomum subsp. goniocalyx to be best classified in Stenotomum Group. Groups Andigenum, Chaucha, Phureja, and Stenotomum are clearly the most unnatural by any phylogenetic criterion. We maintain them as cultivar-groups only because they contain useful characters of ploidy or tuber dormancy mentioned in our keys that provide useful traits for breeders. If different classification needs become useful (such as tuber colors or disease resistances), additional and coexisting cultivar-group classifications can be made, as is allowed by the ICNCP.
We key out but do not provide cultivar-group name(s) for the modern advanced tetraploid varieties of potato (classified previously as S. tuberosum subsp. tuberosum or Group Tuberosum). These modern varieties have resulted from many separate crosses between Andigenum Group, Chilotanum Group, other cultivar-groups, and up to 16 wild species (Ross, 1986
; Plaisted and Hoopes, 1989
; Grun, 1990
). The Chilean landraces and modern varieties differ as a group by isozymes (Ortiz and Huamán, 2001
). These complex hybrid origins provide perhaps one of the strongest arguments for the necessity of the treatment of cultivated potatoes as cultivar-groups, rather than as species. We avoid their simple and traditional classification as Tuberosum Group at this time because they are not the subject of study here, and we think that breeders may benefit from cultivar-groups reflecting their actual use in breeding. For example, breeders typically group potatoes by tuber color and shape reflecting market classes such as long reds, round reds, long whites, round whites, yellows, or russets vs. smooth skins, forming potential cultivar-groups. We would make this classification only after consultation and consensus with user groups.
The ICNCP encourages (but does not require) nomenclatural standards (analogous to types) for cultivars, but no system of typification or use of standards is needed for the cultivar-group names. Cultivar-groups are intended to be solely classifications of convenience based on user-defined needs with no implication of relationships. To our knowledge, standards have never been designated for the cultivars. Many names have been published, however, for the Chilean landraces of "subsp. tuberosum" (Castronovo, 1949
; Kostina, 1978
), modern clones of subsp. tuberosum (Hamester and Hils, 1998
), and landraces in Mexico (Ugent, 1968
), South America (Hawkes, 1944
, 1947
), Peru (Soukup, 1939
; Vargas, 1949
, 1956
; Ochoa, 1958
), and Bolivia (Ballivan and Cevallos Tovar, 1914
; La Barre, 1947
; Ochoa, 1990
).
Key to the landrace cultivar-groups of Solanum tuberosum
Our study documents that some cultivated species have some morphological support, but that these characters represent only typical traits and are not absolutely cultivar-group specific (Fig. 3). Consequently, our key (below) will not consistently separate these cultivar-groups. We include non-morphological characters of reaction to frost, tuber dormancy, daylength adaptation, and ploidy level that are not appropriate for keys of wild plants but are needed here as they are major traits used in the recognition of these cultivar-groups. The qualifier terms "mostly" or "usually" could be used throughout the key but are not used for simplicity.
Descriptions of the landrace cultivar-groups of Solanum tuberosum
As documented above, there are different levels of morphological support for the eight cultivar-groups of S. tuberosum we recognize here, and many characters providing this support are polythetic in nature. The best morphologically supported cultivar-groups are Ajanhuiri Group, Curtilobum Group, Juzepczukii Group, and Chilotanum Group. The Andigenum Group, Chaucha Group, Phureja Group, and Stenotomum Group are primarily distinguished by tuber dormancy (Phureja Group) and ploidy. Consequently, we provide separate descriptions for Ajanhuiri Group, Curtilobum Group, Juzepczukii Group, and Chilotanum Group. We provide a single description for Andigenum Group, Chaucha Group, Phureja Group, and Stenotomum Group as one morphological unit, and the reader is directed to the key for the useful characters for breeders and other users. The group descriptions are followed with lists of some of the well-known cultivar epithets. Some of these are illegitimate, such as ‘Jancko’ (meaning white) and ‘Azul’ (purple), because colors are not allowed in cultivar names. Some epithets appear under two cultivar-groups (as ‘Azul’ in the Curtilobum Group and Juzepczukii Group) that also are not allowed because a cultivar epithet cannot be repeated in a denomination class. These and other errors in nomenclature will have to be corrected in the future.
Solanum tuberosum
Solanum tuberosum is here treated as a denomination class for all cultivated potatoes (Trehane et al., 1995
, p. 68). Plants semi-rosette to ascending to erect, to 0.4–1.4 m tall; stems 5–19 mm wide at base, green to purple or splotched with green and purple, branched; leaves odd-pinnate, diverging from the main stem at about right angles or upright and at an angle of up to 25° from the main stem, terminal leaf tips straight to arched downwards at tip, with 3–8 pairs of lateral leaflets; interstitial leaflets absent or present, with up to 20 pairs; secondary leaflets on the petiolules absent or present, with up to 40 pairs; leaflets with apex acute to acuminate, base oblique, rounded to cuneate to cordate, leaflets ovate to elliptical, nearly glabrous to densely pubescent, margins straight to undulate, petioulate to decurrent; pseudostipular leaves auriculate; inflorescence terminal and lateral; peduncles 3–22 cm long; 4–25 flowers per inflorescence; pedicel 10–35 mm long, articulate very near the top to below the middle; calyx smooth at base or with an encircling horizontal rib below the calyx lobes, regular or irregular with lobes in 1 + 2 + 2 or 2 + 3 groups, tube 3–10 mm long, lobes 1–5 mm long, short and acute to long attenuate, acumens 1–8 mm long; corolla 2–6 cm in diameter, rotate to rotate-pentagonal with short acumens, white to blue to purple to pink, lined or mottled; anthers 3–10 mm long, cordate at base; stigma inserted to exserted up to 7 mm from anther tube; fruits globose to long ovoid, medium to deep green, uniform or with white or purple spots or bands, to purple, 1–4 cm long; tubers with skin color white-cream to yellow to pink to red-purple to purple, uniform throughout or with secondary color in the eyes, eyebrows, around the eyes, stippled or scattered, flesh color white to cream to yellow to orange to red to purple to violet, uniform throughout or with secondary color distributed in the vascular ring or medulla, stippled or scattered, tuber shape globose to ovate to obovate to oblong to elliptic to elongated, smooth to knobby to digitate, tuber eyes shallow to deep, sprouting or dormant at harvest, chromosome number 2n = 2x = 24, 2n = 3x = 36, 2n = 4x = 48, or 2n = 5x = 60.
Landraces distributed throughout the South American Andes to south-central Chile, advanced clones grown worldwide.
Ajanhuiri Group
Plants semi-rosette when young, developing to sub-rosette or to semi-erect, to 0.4–0.7 m tall; stems 8–10 mm wide at base, green to splotched with green and purple, branched; leaves odd pinnate, upright and at an angle of 30–45° from the main stem, terminal leaf tips slightly arched downwards at tip, with 5–6 pairs of lateral leaflets, the uppermost of which are broadly decurrent onto the rachis on the basiscopic side; interstitial leaflets 3–5 pairs, secondary leaflets on the petiolule absent; leaflets with apex distinctly acute, base oblique to rounded, elliptic lanceolate leaflets, densely pubescent on both surfaces, undulate margins; pseudostipular leaves auriculate; peduncle 10–15 cm long; 9–12 flowers per inflorescence; pedicel 21–28 mm long, ratio of length of pedicel from base to articulation/length of pedicel between 0.72 and 0.89; calyx slightly angled, regular, 4–12 mm long, narrowly elliptic lobes shortly acuminate with acumens 1–4 mm long; corolla 2.5–3.5 cm in diameter, rotate-pentagonal, white, white with mauve streaks, blue-mauve, blue-purple; anthers 4–6 mm long; stigma exserted 3–4 mm from anther tube; fruits globose to ovoid, uniformly green or tinged with purple, 2–3 cm long; tubers with skin color white-cream, white with scattered purplish-red, red-violet, purple, flesh color white to cream, uniform throughout, tuber shape ovate to elongated, smooth to knobby, tuber eyes shallow to deep, dormant at harvest, chromosome number 2n = 2x = 24.
Landraces originally distributed in the high Andean altiplano between southern Peru and central Bolivia, at elevations between 3700 and 4100 m a.s.l. However, in Peru only the purple skinned ‘Ajawiri’ is scarcely grown. In the CIP genebank there are 10 cultivars of Ajanhuiri Group. These include ‘Jancko Ajawiri’, ‘Laram Ajawiri’, ‘Jancko Yari’, ‘Wila Yari’, ‘Chañu Yari’, ‘Alka Yari’, and ‘Jancko Sisu Yari’ reported in Huamán, Hawkes and Rowe (1980)
. Others from Bolivia are ‘Chañu Ajawiri’, ‘Wila Palta Yari’, and ‘Wila Anckanche’.
Curtilobum Group
Plants forming a semi-rosette when young, developing to semi-erect and vigorous, to 0.5–0.9 m tall; stems 10–16 mm wide at base, green splotched with purple, branched; leaves odd pinnate, upright and at an angle of 30–40° from the main stem, terminal leaf tips slightly arched downwards at tip, with 5–6 pairs of lateral leaflets; interstitial leaflets 4–6 pairs, secondary leaflets on the petiolule absent; leaflets with apex shortly acuminate, base truncate to rounded to cordate, ovate to elliptical leaflets, sparsely pubescent, undulate to slightly straight margins; pseudostipular leaves auriculate; peduncle 7–8 cm long; 8–14 flowers per inflorescence; pedicel 16–22 mm long, ratio of length of pedicel from base to articulation/length of pedicel between 0.78 and 0.84; calyx smoothly arched, regular, 6–8.5 mm long, elliptic-lanceolate lobes abruptly narrowed at apex to very short pointed acumens 2–3.5 mm long; corolla 3.5–5 cm in diameter, rotate, lilac-purple; anthers 5–6 mm long; stigma exserted 3–4 mm from anther tube; fruits globose to ovoid, green uniform or tinged with purple, 2–3 cm long; tubers with skin color white-cream, white with scattered purple, purple with scattered white, purple, flesh color white, white with scattered purple or purple with scattered white, tuber shape oval-compressed, smooth, tuber eyes shallow to slightly deep, dormant at harvest; chromosome number 2n = 5x = 60.
Landraces originally distributed throughout the highlands above 3800 m a.s.l. from northern Peru to central Bolivia and very rarely in northern Argentina. In the CIP genebank are cultivars mainly differentiated by the tuber skin color and sprout color. These have many different names including ‘Shiri’, ‘Luki’, ‘Waña’, ‘Choquepito’, ‘Mallku’, or ‘Ococuri’, alone or in combination with names describing the tuber skin color like ‘Yuracc’ or ‘Jancko’ (white), ‘Yana’, ‘Laram’, or ‘Azul’ (purple), or ‘Pinta’ (two colored).
Juzepczukii Group
Plants forming a semi-rosette when young, developing to semi-erect, to 0.4–0.8 m tall; stems 10–15 mm wide at base, green to green splotched with purple, branched; leaves odd pinnate, upright and at an angle of 25–60° from the main stem, terminal leaf tips slightly arched downwards at tip to straight, with 5–7 pairs of lateral leaflets, the uppermost of which are slightly decurrent onto the rachis on the basiscopic side; interstitial leaflets 1–4 pairs, secondary leaflets on the petiolule absent; leaflets with apex obtuse to acute, base cuneate or rounded, broadly ovate to broadly elliptical leaflets, rugose, sparsely pubescent; undulate to slightly straight margins; pseudostipular leaves auriculate; peduncle 7–16 cm long; 10–15 flowers per inflorescence; pedicel 22–35 mm long, ratio of length of pedicel from base to articulation/length of pedicel between 0.77 and 0.93; calyx smoothly arched, regular, 4–10 mm long, triangular-lanceolate or elliptic-lanceolate lobes terminated in pointed acumens 2–4.5 mm long; corolla 3–4 cm in diameter, rotate, lilac-purple, dark red-purple, medium to dark purple; anthers 3–5 mm long; stigma exserted 1–2 mm from anther tube; fruits globose to ovoid, green to green tinged with purple, 0.5–1 cm long; tubers with skin color white-cream, white with scattered purple, red with scattered white, purple with scattered white, purple, flesh color white to cream, tuber shape ovoid, oblong or elliptical, tuber eyes shallow to medium deep, dormant at harvest; chromosome number 2n = 3x = 36.
Cultivars originally distributed throughout the highlands above 3800 m from northern Peru to central Bolivia and very scarcely grown in northern Argentina. In the CIP genebank there are 34 different cultivars of Juzepczukii Group including those 21 reported by Schmiediche, Hawkes, and Ochoa (1980)
. These include ‘Jancko Sisu’, ‘Laram Sisu’, and ‘Parco Sisu’ that are putative natural hybrids between Ajanhuiri Group and S. acaule (Johns et al., 1987
) and were described by Ochoa (1990)
. The most common cultivars are ‘Kaisalla’, ‘Kanchillo’, ‘Pariña’, ‘Pechuma’, ‘Pinku’, ‘Piñaza’, ‘Mallku’, ‘Luki’, ‘Shiri’, alone or in combination with the tuber skin color like ‘Yuracc’ or ‘Jancko’ (white); ‘Yana’, ‘Chiar’, ‘Laram’, or ‘Azul’ (Purple), ‘Wila’ (red-purple); ‘Morocc’ (two colored).
Chilotanum Group
Plants ascending to erect, to 0.4–1.0 m tall; stems 6–16 mm wide at base, green or splotched with purple, rarely purple splotched with green, branched; leaves odd pinnate, diverging from the main stem at about right angles or upright and at an angle of up to 50° from the main stem, terminal leaf tips slightly to highly arched downwards at tip, with 3–6 pairs of lateral leaflets; interstitial leaflets absent or present, with up to ten pairs, secondary leaflets on the petiolule generally absent, when present with up to five pairs; leaflets with apex acute to shortly acuminate, base generally cordate, sometimes rounded, rarely truncate or cuneate, ovate to ovate-elliptic to broadly elliptic-lanceolate, nearly glabrous to densely pubescent; generally shiny leaf surface, leaflet margins straight, rarely undulate; pseudostipular leaves auriculate to semielliptic, falcate; flowering absent or scarce under short days, peduncle up to 10 cm long; pedicel 10–20 mm long, ratio of length of pedicel from base to articulation/length of pedicel about 0.50; calyx regular, up to 8 mm long; corolla 2–4 cm in diameter, generally rotate with prominent acumens, white to pale pink, pale blue, or blue-purple to red-purple, uniform or with white acumens; anthers 5–7 mm long, cordate at base; stigma inserted to exserted from anther tube; fruits globose, about 2 cm long; tubers with skin color white-cream to light yellow to pink to red-purple to purple, uniform throughout or with secondary color in the eyes, eyebrows, around the eyes, stippled or scattered, flesh color white, cream, light yellow, rarely red to purple, uniform throughout or with secondary color stippled or scattered, rarely in the vascular ring or medulla, tuber shape globose to ovate to oblong, rarely elongated, generally smooth, tuber eyes generally shallow, rarely deep, dormant at harvest, chromosome number 2n = 4x = 48.
Cultivars originally distributed in the island of Chiloé and adjacent islands in the Chonos Archipelago in Chile. In the CIP genebank there are 143 different accessions of Chilotanum Group native to Chile. Among the most widely distributed are ‘Chapiquina’, ‘Corahila’, ‘Chamizuda’, ‘Clavela’, ‘Azul’, ‘Mantequilla’, ‘Magelanes’, ‘Michune’, ‘Palmeta’, ‘Pichuna’, ‘Cielo’, ‘Chaitenera’, and ‘Camota’. Castronovo (1949)
and Kostina (1978)
described many of these cultivars.
Andigenum Group, Chaucha Group, Phureja Group, and Stenotomum Group
Plants semi-erect, erect, decumbent or prostrate, 0.4–1.4 m tall; stems 5–19 mm wide at base, green to purple, uniform or splotched with purple or green, branched; leaves odd pinnate, upright and diverging from the main stem at an angle of up to 25°, rarely at about right angles, terminal leaf tips straight to slightly arched downwards at tip, with 3–8 pairs of lateral leaflets; interstitial leaflets absent or present, with up to 20 pairs, secondary leaflets on the petiolule absent or present, with up to 40 pairs; leaflets with apex acute to acuminate, base oblique, rounded to cuneate to cordate, leaflet shape ovate to elliptical, apex acute to acuminate, base cordate to attenuate, leaf surface dull to shiny, nearly glabrous to densely pubescent; leaflet margins straight to undulate, petiolulate; pseudostipular leaves auriculate; peduncle 3–22 cm long; 4–25 flowers per inflorescence; pedicel 10–35 mm long, ratio of length of pedicel from base to articulation/length of pedicel between 0.34 and 0.84; calyx smoothly arched at base to greatly angled and ribbed, regular or irregular with lobes in 1 + 2 + 2 or 2 + 3 groups, 3–10 mm long, short and acute to long attenuate, acumens 1–8 mm long; corolla 2–6 cm in diameter, very rotate to rotate-pentagonal, white to lilac to pink to blue to purple, uniform or with a secondary color stippled, in bands, in the star, or white acumens in the adaxial, abaxial, or both sides; anthers 3–8 mm long, cordate at base; stigma inserted to exserted 7 mm from anther tube; fruits globose to ovoid, green, uniform, or tinged with white or purple spots or bands, 1–4 cm long; tubers with skin color white-cream to yellow to pink to red-purple to purple, uniform throughout or with secondary color in the eyes, eyebrows, around the eyes, stippled or scattered, flesh color white to cream to yellow to orange to red to purple to violet, uniform throughout or with secondary color distributed in the vascular ring or medulla, stippled or scattered, tuber shape globose to ovate to obovate to oblong to elliptic to elongated, smooth to knobby to digitate, tuber eyes shallow to very deep, sprouting or dormant at harvest, chromosome number 2n = 2x = 24, 2n = 3x = 36, 2n = 4x = 48.
Cultivars originally distributed in the highlands of Mexico, Guatemala, Venezuela, Colombia, Ecuador, Peru, Bolivia, and northern Argentina. The characters distinguishing Andigenum Group, Chaucha Group, Phureja Group, and Stenotomum Group are shown in the key above.
In the CIP genebank, out of the 3227 accessions classified in Andigenum Group, 2379 have been found to be morphologically different and/or have different isozyme patterns (Huamán, Ortiz, and Gomez, 2000
; Huamán et al., 2000
). About half of the remaining 848 accessions could be different. The geographical distribution includes the highlands of Mexico, Guatemala, Venezuela, Colombia, Ecuador, Peru, Bolivia, and northern Argentina, at elevations between 1000 and 4300 m a.s.l. Widely known cultivars in Peru are ‘Ccompis’, ‘Yana Imilla’, ‘Yuracc Imilla’, ‘Huagalina’, ‘Alka Tarma’, ‘Hualash’, ‘Cusi’, ‘Bole’,
