Before we can begin to discuss speciation, we must first attempt to define a species. Commonly, populations that can be distinguished by prominent
5
© J.F. Hancock 2004. Plant Evolution and the Origin of Crop Species,
2nd edn (J.F. Hancock) 100
morphological differences are considered to be species. This ‘taxonomic species concept’ is often successful in identifying separately evolving groups, but in some instances can lead to artificial groupings. Quite distinct morphotypes can retain the ability to freely interbreed, making them in real- ity one genetic entity; and large differences in morphology are sometimes influenced by only a few genes, which do not necessarily reflect the diver- gence of the whole genome (Chapter 1). These ambiguities have led to fre- quent revisions of important crop assemblages, such as oats, rice, wheat and sorghum, as different taxonomists have reviewed the available data on natural populations.
Another way to distinguish species is by directly testing their capacity to successfully reproduce. Mayr (1942) defined the ‘biological species’ as
‘groups of actually or potentially interbreeding natural populations, which are reproductively isolated from each other’. If taxa are reproductively iso- lated, they are evolving separately and therefore must have an identity of their own. This concept has gained widespread approval and is currently the most popular (Howard and Berlocher, 1998; Schemske, 2000).
While the biological species concept allows for the unambiguous delin- eation of species, it is still not without occasional problems. Strongly diver- gent groups of plants often maintain some degree of interfertility even though they differ at numerous loci and are effectively evolving on their own. As we shall discuss below, sunflower and violets provide particularly striking examples. Plants also show great ranges in fertility from obligate out- crossing to complete selfing to apomixis (uniparental). In a highly inbred or apomictic group, every individual would be a species according to the bio- logical species concept. As we have already mentioned, many of the grain and legume species are highly inbred, and most of the starchy staples, such as banana, cassava, potato, sugar cane, sweet potato, taro and yam, are only propagated through asexual means.
Harlan and deWet (1971) developed the ‘gene pool system’ to deal with varying levels of interfertility between related taxa (Fig. 5.1). They recog- nized three types of genic assemblages:
1. Primary gene pool (GP-1) – hybridization easy, hybrids generally fertile.
2. Secondary gene pool (GP-2) – hybridization possible but difficult, hybrids weak with low fertility.
3. Tertiary gene pool (GP-3) – hybrids lethal or completely sterile.
The primary gene pool is directly equivalent to the biological species.
The recognition of GP-2 and GP-3 allows other levels of interfertility to be incorporated into the overall concept of species. These are related taxa which share a considerable amount of genetic homology with GP-1, but are divergent enough to have greatly reduced interfertility. Several agronomi- cally important groups have been described using this system, including legumes (Smartt, 1984), wheat (Fig. 5.2) and most of the other cereals (Table 5.1).
Simpson (1961) developed the idea of an ‘evolutionary species’ to mini- mize the problems associated with uniparental species. He suggested that a species must meet four criteria: (i) is a lineage; (ii) evolved separately from other lineages; (iii) has its own particular niche or habitat; and (iv) has its own evolutionary tendencies. This definition fits uniparental species better than the biological species concept, but we are still left to decide on what constitutes a lineage. Templeton (1989) expanded this theme by using molecular data to construct phylogenies in his ‘cohesion species concept’. He defined a species as an ‘evolutionary lineage, with the lineage boundaries arising from the forces that create reproductive communities (i.e. cohesive mechanisms)’.
Numerous other concepts have been developed to include ecological with reproductive criteria in defining species (Levin, 2000; Schemske, 2000).
Nevertheless, it is clear that no model solves all of the potential problems in trying to define separate evolutionary units. Each has its own strengths and weaknesses. Levin (2000) suggests that ‘the choice of a species concept has to do, in part, with the perspective that gives one satisfaction’. Probably the most definitive definition is the concept of biological species since it can be
Hybrids with GP-1 anomalous, lethal or completely sterile
All species that can be crossed with GP-1 with at least
some fertility in F1s GP-1 Subspecies A:
cultivated races
Subspecies B:
spontaneous races GP-1
Gene transfer not possible or requiring radical techniques
Gene transfer possible but may be difficult BIOLOGICAL SPECIES GP-2
GP-3 GP-2 GP-3
Fig. 5.1. Schematic diagram of primary gene pool (GP-1), secondary gene pool (GP-2), and tertiary gene pool (GP-3) (used with permission from J. Harlan, © 1975, Plants and Man, American Society of Agronomy, Madison, Wisconsin).
directly tested – two individuals can either successfully reproduce or they can- not. Of course, environmental and genotypic variation can cloud even this approach, but it does minimize the number of subjective judgements. As a rule of thumb, most evolutionists consider species to be those groups that are reproductively isolated, although plant scientists are more willing to accept some degree of hybridization between otherwise distinct species.