Taxonomic and Ecogeographic Studies of the Latin-American Cucurbitaceae: Results of an Inter-institutional Effort between the International Board for Plant Genetic Resources and the National Herbarium of Mexico

Rafael Lira Saade

International Board for Plant Genetic Resources
Via della Sette Chiese 142, 00145 Rome, Italy


The Curcurbitaceae family includes approximately 80 genera and 900 species. This family represents one of the most important plant groups, because it has many species that have been used as food or for other purposes for hundreds of years. Among the genera that are most important from this point of view, there are four of American origin: Cucurbita with five cultivated species (C. argyrosperma, C. moschata, C. ficifolia, C. maxima and C. pepo); Sechium with two cultivated species (S. edule, and S. tacaco); Sicana with one cultivated species (S. odorifera); and Cyclanthera, a large genus that includes three cultivated and/or semicultivated species, ( C. pedata).

Despite the importance of these genera, the systematic study of at least Sicana and Sechium had not been carried out, or at the most, as in Cyclanthera, only some details are already known. In contrast for the genus Cucurbita, a lot of information has been generated and published at different times and places, so that the summarization and analysis of it seems necessary. Additionally, there are no modern taxonomic treatments of these genera, including the genus Cucurbita, that could be used for an easy and correct identification and assessment of the situation of the taxa in the field. Also, it was not really known which of them were the most important ones and if they were really managed for their conservation, study and use as genetic resources.

Considering the above statements, in 1990 the International Board for Plant Genetic Resources (IBPGR) and the National Herbarium at the Institute of Biology of the National University of Mexico (MEXUIBUNAM) initiated a project of which the main objective was to summarize and/or enlarge the information concerning these topics.

The purpose of this paper is to present some of the most important results of this inter-institutional collaborative project, which include the rediscovery of a species formerly considered extinct, the discovery of a new one, and advances in the knowledge of the interspecific relationships and conservation needs of some of the four genera involved. These results were obtained through field work, collection studies, and the use of different biosystematic techniques, whose implementation would not be possible without the support of an institution such as MEXUIBUNAM which possesses important infrastructure, as for example, a herbarium (the most important collection of Mexican plants), a botanical garden, as well as laboratory and greenhouse facilities.


First of all I would like to thank the organizers of this meeting for giving me the opportunity of sharing with all of you the results of this inter-institutional project between the International Board for Plant Genetic Resources and the National Herbarium of Mexico, that deals mainly with the taxonomy and the ecogeographic distribution of economic Latin American Cucurbitaceae, as a basis for their future use and conservation.


The Cucurbitaceae family includes approximately 80 genera and 900 species. This family represents one of the most important plant groups, because of having many species that are used as food or for other purposes. Many of its species have been 'generated' by humanity using complex processes of domestication, and they have ben part of the human diet as vegetables and as oil and protein sources for thousands of years.

Among the genera that are most important from this point of view, there are four of American origin: Cucurbita with five cultivated species, Sechium with two cultivated species, Cyclanthera, a large genus that includes three cultivated and/or semi-cultivated species; and Sicana with one cultivated species.

Even given the great importance of these genera, the systematic study of at least Sicana and Sechium had not been carried out, or at the most, as in Cyclanthera, only some details were already known. In contrast, for the genus Cucurbita, a lot of information has been generated and published at different times and places, so that the summarization and analysis of it seems like a necessary work to do. Additionally, in any case, including the genus Cucurbita, there was not any modern taxonomic treatment of these genera that could be used for an easy and correct identification of the taxa. Also, it was not really known which of them were the most important ones and whether they were really managed for their conservation, study and use as genetic resources.

Considering the above statements, in 1990 the IBPGR and the National Herbarium at the National University of Mexico initiated a project of which the main purpose was to summarize, or enlarge, the information concerning the topics already mentioned.

The general objectives to be accomplished in this project were:

Methodology and Sources of Information

In order to carry out this work, many sources of information and methodologies have been used, including bibliographic revision, collection studies, field work (mainly in Mexico and Central America), as well as some laboratory work.


The outstanding results for each genus can be summarized as follows:


The first genus that I will be talking about is Cucurbita. As was mentioned before, it includes five cultivated species widely used as food The bibliographic revision revealed that the use of electrophoretic techniques (mainly for isozymes), as well as hybridization studies, have helped to obtain important advances in the taxonomic knowledge of this genus, especially with respect to the relationship between the five cultivated species and the closest wild ones. At the present it is recognized that Cucurbita has 10 to 12 species or groups of species, five of which represent the cultivated species and their wild ancestors.

The relationship between 'wild' and 'cultivated' is at the present clear for three of the five cultivated species. C. argyosperma, a very important traditiona1 crop cultivated from the southwestern United States to Central America, whose wild ancestor is C. sororia distributed from Mexico to Central America; C pepo may be the most important cultivated species of the genus, it is very closely related to C. fraterna and C. texana; and C. maxima, a South American native crop, but at the present cultivated all over the world, whose wild closest relative is C. andreana, an endemic species from Argentina and Uruguay. Meanwhile for the two other species, C. moschata and C. ficifolia, it has not been possible to define the closest wild species. C. ficifolia is a relatively homogeneous species that does not produce totally fertile hybrids with any other species of the genus; in contrast, C. moschata is a very variable crop that forms hybrids relatively easily with several other species of the genus.

Based on the fact that most of the information used in the definition of the relationships formerly mentioned has been obtained by studies of experimental hybridization, we realized the need to test for the existence of interspecific gene flow in natural conditions. As a first step this has been accomplished for C. fraterna, an endemic species of northern Mexico in the states of Tamaulipas and Nuevo Léon and one of the closest wild relatives of C. pepo. This study has been carried out in a small area of Tamaulipas State, not too far from the Ocampo caves, one of the most important archeological sites for C. pepo.

So far, the results obtained have shown that C. fraterna is more common than was thought before from previous reports. It has also been interesting to find that this species is more common as a weed in cultivated fields, than in other kinds of habitats typically inhabited by many wild species of Curcurbita, such as roadsides or abandoned fields. An important finding is that the most common cultivated species in the studied area, and the ones that at the present are represented by primitive landraces, are C. moschata and C. argyrosperma, and not C pepo, whose cultivation is less common and is only represented by one commercial cultivar. Nevertheless, the finding of sweet fruits in some populations of C fraterna (which is not a common feature in wild species), the obtaining of some fruits due to in situ artificial crossings between C. fraterna and C. argyosperma and the ethnobotanical information from the native peasants of the area, indicating that sometimes the fruits of the cultivated species have a bitter flavour, are strong evidence suggesting the existence of gene flow between C. fraterna and the two more commonly cultivated species in Tamaulipas. The evaluation of hybrid progenies in greenhouses and through an isozyme study using electrophoretic techniques, have enhanced the detection of the genetic infiltration confirming the former assumptions.

Taking into consideration the fact that C fraterna and C. agyrosperma are not closely related, as has been demonstrated by biosystematic studies, this result is very important, because besides relatively changing the relationships scheme in Cucurbita, it also permits the opening of new perspectives in which we should attempt to evaluate C fraterna in detail to be used as a probable source for the improvement of at last three of the cultivated species. One of them is obviously C. pepo, which until now has not ben known to be resistant to many virus diseases, and the other two are C. argyrosperma and (probably) the closely related C. moschata.

The above information should be considered important in the context of the in situ conservation strategies of traditional agriculture systems, which, unfortunately in my opinion, until now have been a sort of myth for the agencies and foundations that support conservation activities. Traditionally, the Cucurbita germplasm has been protected in frozen banks, where only part of the genetic diversity can be conserved over a long term period, and where the natural evolution of wild taxa interacting with their cultivated relatives will hardly take place.

With respect to the germplasm collections of Cucurbita, according to the bibliographic information concerning this topic and the visit to institutions in Mexico and Costa Rica, we were able to observe that many collections of both wild and cultivated species have shown resistance to important diseases affecting the five cultivated ones. However, we found that most of the species of this genus, especially the wild ones, are very poorly represented in these collections.

On the other hand the revision of material, and mainly the intensive field work carried out, has given us better knowledge about the morphological variation, distribution and ethnobotanical aspects of the species of Cucurbita, mainly of the less known of them. As an example, let me point out that we have been able to collect material such as herbarium specimens, seeds, and even roots of Cucurbita pedatifolia, C. scabridifolia and C. radicans, all of them very poorly represented in the Latin American germoplasm banks, and rarely present in similar collections in other parts of the world. Finally, with respect to the laboratory work: on Cucurbita, we have been preparing an identification manual for seeds, both carbonised and intact ones, that will be used for the identification of archeological remains.


The second genus to be dealt with is Sicana; this genus includes only three species. The most important one is S. odorifera, which is cultivated all over Latin America. The other two are: S. sphaerica, a wild species only known from Jamaica, and S. trinitensis, a wild species from Trinidad and Tobago, Venezuela and maybe Bolivia in South America. The bibliographic and specimens revision revealed a complete lack of knowledge about the genus, as well as about the importance and relationships of the cultivated species. Two instances of this are that many floras describe this genus as monospecific, and that all the ethnobotanical information we have about the cultivated species is that its mature fruits are used to prepare candies and, because its smell, are used as a sort of natural air cleaner, while the young ones are occasionally mentioned as being used as a vegetable.

In order to cover some deficiencies in our knowledge of the genus, and particularly about the cultivated species, a morphological study of the three species was carried out: field collections were done in Mexico and Central America that enhanced a better documentation of the cultivated species; for the first time chromosome counts were done for Sicana odorifera, and the first chemical analysis of its seeds was made.

The results of these studies showed that Sicana odorifera is a much more variable crop than we thought, that its diploid number is 2n=40, and that at least from the morphological point of view S. trinitensis might be its probable wild ancestor. However, the most important result was the finding that the big seeds of S.odorifera contain 39% of oil and more than 25% of protein, which along with the fact that each fruit can produce up to 900 seeds, suggests that this species might be a potential source of these important nutrients. Unfortunately the representation of Sicana in the germplasm banks is very poor. The wild species are not represented at all, while the cultivated species are represented by only 13 collections, including the ones collected by our project. Also it was found that the two wild species apparently have never been redocumented, not even by herbarium specimens, since they were originally described the first half of this century.


The next genus to be dealt with is Sechium, a genus that includes two cultivated species: S. edule is the most important species and at the present time is cultivated in many different parts of the world because of its edible fruits and roots, while S. tacaco is an endemic crop only cultivated in Costa Rica, where its small fruits are a very important and traditional table vegetable.

At the beginning of the project, it was found that there were two different and contrasting taxonomic conceptions of this genus. The first one proposed by Charles Jeffrey in 1978 included the two cultivated species and five wild ones, while the second approach, proposed by Linda Newstrom in l986, accepted only three species, the cultivated S. edule and two wild species.

With the purpose of obtaining information about this matter, collections of all the species involved in the problem were made and a lot of specimens, including all the types, were revised. Two of the most important results of this part of the work were: the re-discovery of S. hintonii, a wild species only know from the type collection collected in 1933-35, and that at the present we only know lives in the states of Mexico and Guerrero, and the discovery of a new species, S. chinantlense in the north of the state of Oaxaca, Mexico, whose stamens are quite different from those normally found in the rest of the species of Sechium. The importance of these findings is that the two species are closely related to Sechium edule. In addition to these interesting collections of wild species, an amazing morphological and phenological variation of the cultivated species S. edule could be documented in southern Mexico, especially in the states of Oaxaca and Chiapas, confirming that this area is a part of the most important centre of diversity for this crop.

With all the material collected, a palynological study of all the species of the genus, and some other species belonging to the same subtribe Sicyniae was carried out. The results showed homogeneous pollen for all the species recognised as Sechium, which confirmed the former assumptions, based on gross morphology, habitat, etc., that indicated a large number of species for the genus. Simultaneously, chromosome counts for most of the species were done. Some preliminary results indicated the existence of interspecific differences and mainly with respect to the reports of the species of other genera of the same subtribe.

Presently we are initiating some chemiotaxonomic and molecular studies, as well as some work dealing with the reproductive biology of the new wild species S. chinantlense, and its possible gene flow with the different types of wild and cultivated S. edule that grow in the north of the state of Oaxaca. The results obtained in this study are in contrast to those mentioned for C. fraterna, with respect to its relationships with the cultivated species. The ethnobotanical evidence concerning the gene flow between S. chinantlense and S. edule is not very consistent, and the biosystematic evidence that has been obtained is not definite, mainly because of the lack of features in the wild and cultivated plants that suggest any kind of genetic infiltration. In addition, only four fruits could be obtained after carrying out numerous crosses between S. chinantlense and cultivated and wild forms of S. edule. However, the grown plants were not normal and died some time after germination, confirming that, besides the morphogical differences between S. edule and S. chinantlense, the latter is clearly a genetically different species.

Considering all of these results, at the present time I think that Sechium includes 10 species, the 2 cultivated ones and 8 wild species naturally distributed within relatively small areas from central and southern Mexico to Panama. The very narrow distribution pattern of these species, the disturbance of their habitats, as well as the fact that only two of them thrive in protected areas, are strong arguments for considering all of these species as endangered. Regarding the ex situ conservation of the two cultivated species of Sechium, we found that only S. edule has been protected by this method. However during 1990 I personally verified the total loss of the most important gene bank of S. edule in the world in Costa Rica at the Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), and recently I have heard that the Mexican collection located in Huatusco, Veracruz, might be lost very soon. In spite of these events, recently I have been informed that in Costa Rica a project, including private and official institutions, is taking place in order to reinstate a germplasm bank of Sechium.

Many management difficulties and the high costs due to the need to have plantations of these plants, are the main arguments used to justify the loss of the collections of Sechium, and consequently it is necessary to think of other forms of conservation. Considering the facility that these plants have for reproducing vegetatively, is possible to think that in vitro conservation or micropropagation could be the alternative. There are already some advances to this in respect with S. edule, and probably it can also be done with the wild species. However, I think that there is a serious risk of the genetic resources of Sechium (especially the wild ones) being lost while this happens.


The last genus that I will talk to you about is Cyclanthera. This genus has three useful species whose fruits are common table vegetables in several countries of Central and South America. The most important species is C. pedata, a cultivated species which at present is used in several countries of the world. The other two species are C. brachystachia and C. brachybotrys; these are semi-cultivated only in some regions of South America. In addition Cyclanthera includes numerous wild species, but the definition of their relationships deserves a modern taxonomic revision. However, Cyclanthera was too large to be worked on during this project, and consequently I have just summarised the information concerning the important species, as well as proposing some research activities for studying this interesting genus.

Publications, Talks and Monographs

Most of these results have already been published in several journals, or in some cases presented in conferences, and at the present the final details of the monograph are being decided. This monograph will be published as a part of the IBPGR's series called 'Systematic and Ecogeographic Studies on Crop Gene Pools'. Besides all the data about the importance of the species, it will include taxonomic treatments of the four genera, with information about the scientific and folk nomenclature, typification, descriptions, identification keys, distribution maps, and illustrations showing different aspects of the genetic diversity of the species. It is important to say that, in addition to the main results of the project, we have collated a bibliographic database with more than 3000 references about the four genera studied and several topics related to their importance as genetic resources, as well as a taxonomic and ecogeographic one that at present contains more than 2000 records of the species studied.

Perspectives of the Study

Even though some aspects have been solved in the last three years, it is also true that there is still a lot to be done. The search for the wild ancestors of the two cultivated species of Cucurbita that are still unknown, the complete study of Cyclanthera and the more in-depth study of the variation and intrageneric relationships of Sechium using molecular biology techniques, seem the work to be accomplished in the near future. In addition, in the long term, some other aspects have to be studied, mainly in order to continue the collection and evaluation of wild taxa as genetic resources for the improvement of the cultivated species, or for the utilization of both the wild and cultivated ones as sources for products such as, oil, proteins, fuel, etc. However to reach these goals, it will be necessary to have a multi-disciplinary project in which many different experts, such as agronomists, taxonomists, chemists, phytopathologists, etc, work in cooperation. I am sure that behind this diversity of forms and colours, an amazing genetic diversity is hidden, and that we still have the opportunity to discover, study, conserve and use it.

Preface  |  Contents List  |  Congress Report  |  Workshop Conclusions  |  List of Authors