The criterion for evaluating the differences between populations of a single species is based essentially on the gene frequency differences. When these differences are many and it is advantageous to consider populations as separate entities, they are classified into races. These gene-frequency changes usually result from the response of a population to the selective forces operating within a particular environment. The race-formation is accelerated by barriers that reduce gene exchange between populations. The potential for gene exchange, however, enables all these different populations to be considered as members of a single species. It is only when populations have achieved sufficient differences to inhibit any gene exchange at all between them that they are considered to have diverged sufficiently to have reached the level of separate species.

Species
A species is an interbreeding group distinct from such other groups. The transition of racial differences to species differences is marked by a qualitative change accompanied by reproductive separation or isolation. Mechanisms preventing gene exchanges are termed isolating mechanisms. They include all kinds of factors and according to some, even the geographical and spatial isolation. Such geographically separated populations called allopatric populations do not have the opportunity for gene exchange. However, it is debatable whether they remain reproductively isolated, if given opportunity. Therefore, isolating mechanisms, it is argued, should be restricted to include only those factors which prevent gene exchange between populations of the same locality i.e. sympatric populations. According to Mayr (1963) two types of isolating mechanisms are available : premating and postmating. The former includes seasonal or habitat isolation, behavioural isolation and mechanical isolation while gametic and zygotic mortality and hybrid sterility involve postmating isolating mechanisms.

In general, however, the barriers separating the species are not confined to a single mechanism. The Drosophila species D. pseudoobscura and D. persimilis are isolated by habitat, courtship time and mating behaviour. In plants, both pre- and postmating isolating mechanisms may be operative in the same interspecific cross, the postmating barriers functioning when the former have been bypassed. For example, it has been shown that the cross Gilia australis (O) x G. splendens (O) usually fails because of retarded growth of australis pollen tubes. On the other hand, sple-idens pollen tubes may reach australis ovules in reciprocal cross but the hybrid embryo, however, fails to develop because of degeneration of the endosperm.

Modes of speciation
It. has been seen that speciation should occur in the following sequence : (i) genetic differentiation among a sympatric population, (ii) overlap of differentiated populations in a sympatric area, and (iii) selection for sexual isolating mechanisms. Demonstration of this sequence among natural populations has been attempted by comparing the degree of sexual isolation between different sympatric and allopatric populations; sexual isolation being strongest among the sympatric populations. Grant (1963) reported, that of the nine species in Gilia, those that are most difficult to cross are the sympatric ones. The allopatric species, show no barriers against hybridization, although all the crosses give sterile progeny. Dobzhansky and co-workers (1964) found five morphologically identical races of Drosophila paulistorum in Central and South America that show complete isolation from each other when they are found in same locality. They still can exchange genes with various strains of a 'transitional race' and therefore, can not be considered as full-fledged species. Instead, they are called 'incipient species' or 'semispecies' and D. paulistorum is called superspecies. The genetic basis for sexual isolation between these populations was found to be polygenic, located on all the chromosomes (Ehrman, 1960).

In plants, fertile hybrids can introduce genes from one species into the other, a phenomenon called 'introgressive hybridization' by Anderson (1949). Mangelsdrof (1974) believed that the introduction of genes from Tripsacum into Zea mays must have occurred by repeated backcrosses. Polyploidy may also enable sterile hybrids to produce fertile gametes (allpolyploidy) and since these gametes are diploid relative to the haploid gametes of the parental species, a new species is born in one step. The sequence of evolutionary events in speciation, therefore, seems to start with race formation and end with reproductive isolation brought about by spatial or geographical separation and sexual isolating mechanisms. The later, in turn, could be achieved accidently or by selection against hybrids.