The Cucurbitaceae of India: Accepted names, synonyms, geographic distribution, and information on images and DNA sequences

Abstract The most recent critical checklists of the Cucurbitaceae of India are 30 years old. Since then, botanical exploration, online availability of specimen images and taxonomic literature, and molecular-phylogenetic studies have led to modified taxon boundaries and geographic ranges. We present a checklist of the Cucurbitaceae of India that treats 400 relevant names and provides information on the collecting locations and herbaria for all types. We accept 94 species (10 of them endemic) in 31 genera. For accepted species, we provide their geographic distribution inside and outside India, links to online images of herbarium or living specimens, and information on publicly available DNA sequences to highlight gaps in the current understanding of Indian cucurbit diversity. Of the 94 species, 79% have DNA sequences in GenBank, albeit rarely from Indian material. The most species-rich genera are Trichosanthes with 22 species, Cucumis with 11 (all but two wild), Momordica with 8, and Zehneria with 5. From an evolutionary point of view, India is of special interest because it harbors a wide range of lineages, many of them relatively old and phylogenetically isolated. Phytogeographically, the north eastern and peninsular regions are richest in species, while the Jammu Kashmir and Himachal regions have few Cucurbitaceae. Our checklist probably underestimates the true diversity of Indian Cucurbitaceae, but should help focus efforts towards the least known species and regions.


Introduction
Jeff rey's (1980) and Chakravarty's (1982) checklists of the Cucurbitaceae of India are now more than three decades old. Over this time, knowledge of the family's representatives on the Indian continent has grown considerably through botanical exploration, the additions of Naithani (1990), new treatments for Th ailand (De Wilde and Duyfj es, 2008a) and China , and revisionary work on genera, such as Trichosanthes (De Boer and Th ulin, 2012) and Coccinia (Holstein, in press). Added to this, the online availability of taxonomic literature and specimen images, and molecularphylogenetic studies clarifying natural clade boundaries (e.g., Schaefer et al., 2009;Sebastian et al., 2011;, have led to many taxonomic and nomenclatural changes. Updating the two checklists of Indian Cucurbitaceae was therefore timely, especially since the Cucurbitaceae include several of the World's most important vegetables, such as melon (Cucumis melo), cucumber (C. sativus), watermelon (Citrullus lanatus), pumpkin and squash (Cucurbita spp.), and bitter gourd (Momordica charantia). Having a current list that is linked with molecular data and images may help focus phylogenetic and fl oristic research on undercollected species, and potentially strengthen conservation eff orts.
Here we present a checklist of the Cucurbitaceae of India that treats just over 400 relevant taxon names. For each accepted species, we provide (i) type information including collecting location and herbaria, (ii) synonyms and their types, (iii) information on geographic range inside and outside India, (iv) links to online images of herbarium or living specimens, and (v) brief information on whether or not DNA sequences are available in GenBank at the National Center for Biological Information (http://www.ncbi.nlm.nih.gov), with citation of relevant studies. DNA sequences today are essential; they help in the quick identifi cation of sterile material via characteristic sequence motifs or "barcoding" (an Asia-focussed example is Li et al., 2011) and are required for evolutionary and biogeographic studies (e.g., Sebastian et al., 2011. Even DNA sequences not coming from Indian material can help place the Indian species in context and to recognize if Indian material diff ers from African or Chinese material going by the same name.

Materials and methods
Names that have been applied to Indian Cucurbitaceae were taken from Jeff rey (1980,1981), Chakravarty (1982), and an unpublished compilation provided by Peter Raven (the Missouri Botanical Garden, St. Louis) and Kanchi Gandhi (Harvard University Herbaria, Boston). We also checked fl oras of neighboring or near-by countries, especially Naithani (1990), the Flora of China treatment , and numerous publications by De Wilde and Duyfj es (cited in our reference list). Information on the types (collector and location) of the 400 names was obtained from protologues, most of them available online. For nomenclatural types from India, we updated the state in which the respective specimen was collected to agree with modern administrative units. Taxonomic or nomenclatural synonyms were obtained by checking relevant post-1980 treatments (cited under the respective genus or species).
Distributions within India (by state) and outside India (by country or continent) were taken mostly from Chakravarty (1946Chakravarty ( , 1959Chakravarty ( , 1982, up-dated from fl oristic treatments, such as Lu et al. (2011) and the work of De Wilde and Duyfj es (e.g., 2004a, b, 2006a, b, c, 2007a, b, 2008a, 2010. Th e links to images lead to type specimen images from various herbaria or the efl oraofi ndia website (https://sites.google. com/site/efl oraofi ndia/). Th is website has been created for documenting the fl ora of India and currently has a database of 7500 species and over one million pictures at its e-group links. For each accepted species or relevant synonyms we checked GenBank (http://www. ncbi.nlm.nih.gov) for sequences and the published studies they are related to.

Results and discussion
Comparison with the two 1980s checklists and main causes of name changes Applying recent taxonomic changes resulted in the acceptance of 94 species. Th is is almost unchanged from the species number listed in previous checklists (Jeff rey, 1980: 90 species;Chakravarty, 1982: 100 species). A species no longer included is Zehneria wallichii from central Myanmar. Newly added species include Trichosanthes khasiana and T. quinquangulata. Compared to 1980, generic concepts have changed considerably, with many species names having been moved, especially in the genera Cucumis and Zehneria, and formerly monotypic genera having been merged . Genera no longer accepted are Biswarea (=Herpetospermum), Cucumella (= Cucumis), Dicoelospermum (= Cucumis), Edgaria (=Herpetospermum), Gymnopetalum (=Trichosanthes), Mukia (= Cucumis), Neoluff a (= Siraitia), Praecitrullus (= Benincasa), and Sechium (= Sicyos). All these changes are based on molecular-phylogenetic results, cited under the respective species. Melothria in its modern circumscription is confi ned to the New World and does not occur in India. Its two Indian species have been moved to Cucumis and Solena. Compared to other tropical regions of the size of India, for example, Brazil, the addition of new species records over the past 30 years has lagged behind. We suspect that many species new for India are awaiting discovery in the fi eld and in yet unidentifi ed herbarium material. Since Indian herbaria are reluctant to send out loans, their material probably is understudied.
While 86 native species, including just ten endemics, may not be large numbers, India harbors an exceptional range of tribes as seen in Fig. 1, which shows the placement of the native Indian genera on a Cucurbitaceae family tree with the family's current tribal classifi cation . Many of the Indian species, such as Actinostemma, Gynostemma, Hemsleya, Indofevillea, Momordica and Siraitia belong to old and phylogenetically isolated lineages. Th is is known because 79% of the Cucurbitaceae species occurring in India have been sequenced for one or more genetic markers. Cucumber and melon, which originate in India, both have had their genomes completely sequenced (Huang et al., 2009;García-Mas et al., 2012), and many have been included in family-wide phylogenetic analyses Schaefer et al., 2009;Schaefer and Renner, 2011b). Th e currently 20 species without any DNA sequences in GenBank may be found by searching our checklist for "no published sequences available."

Floristic distribution within India and disjunctions between Africa and India
Th e highest number of species is known from the northeast and peninsular India (Kerala, Karnataka, Tamil Nadu, Andhra Pradesh), the lowest from the Jammu Kashmir and Himachal regions of Western Himalaya. Especially interesting from a phytogeographic standpoint are species ranging from Africa to India, such as Coccinia grandis, Blastania cerasiformis, Corallocarpus conocarpus, C. epigaeus, C. schimperi, Cucumis prophetarum, Dactyliandra welwitschii, Luff a echinata, Momordica cymbalaria, and Zehneria thwaitesii. Th e genera Diplocyclos and Kedrostis also both have species in East Africa and India, but apparently not individual species spanning both continents. Th ese disjunctions would be interesting to study with molecular methods, which might allow inferring arrival times in India.

Conclusion
One of the great technical advances of recent years that are positively aff ecting taxonomy is the easy exchange of photos. Even simple snap shots of living plants (and cer- tainly type images) greatly facilitate deciding the identity of a particular plant, and we hope that our links to the efl oraofi ndia (https://sites.google.com/site/efl oraofi ndia/) will proof useful. Th e greatest caveat concerning our checklist is that the geographic range information inside India is not directly based on specimens, but is more or less copied from Chakravarty (1982) and thus surely incomplete. It is to be hoped that the digization of Indian material in the future will help achieve a deeper study of the Cucurbitaceae of India.  Volz and Renner (2009), e.g., EU683747, EU683740. Comment: Th is was treated as B. dioica Jacq. by Chakravarty (1982), but that species does not occur as far east as India, ranging instead from Spain south to Algeria and Morocco, Sardinia, Corsica, and the Greek Peninsula and east to mid-Poland; a distribution map with all species of Bryonia is provided by Volz and Renner (2009 Efl oraofi ndia at https://sites.google.com/site/efl oraofi ndia/species/a---l/cl/cucurbitaceae/corallocarpus/corallocarpus-epigaeus GenBank: AM981182 from an unpublished paper. Comments: Th e species is used as an anthelmintic (Chopra et al., 1956).  (Naithani, 1990), Madhya Pradesh, Maharashtra, Rajasthan, Tamil Nadu, Tripura, Uttar Pradesh, Uttarakhand. Distribution outside India: Native to Central America. Image: efl oraofi ndia at https://sites.google.com/site/efl oraofi ndia/species/a---l/cl/cucurbitaceae/cucurbita/cucurbita-maxima GenBank: Numerous sequences from the three plant organellar genomes. Comments: Winter squash is cultivated throughout India.