The Halepensia are Mediterranean and Asian and are distinguished by their long creeping rhizomes. They are represented in Pakistan by Sorghum halepense. The Arundinacea are African, but with two possible exceptions, Sorghum stapfii and Sorghum pugionifolium, which are both exclusively Indian.

The Arundinacea are taxonomically the most difficult group of Sorghums. Snowden (1936, 1955) subdivided them into 28 cultivated and 24 related wild “species”, but it is now generally agreed that his cultivated species are best regarded as cultivars. DeWet & Huckaby (in Evolution 21:787-802. 1967) pointed out that since there are no genetic barriers between any of the taxa, wild or cultivated, they should be treated as a single species. This view is perhaps a little extreme and there are sound practical reasons for maintaining the cultivated Sorghums as a separate species from the wild ones. Indeed, it seems best to regard the Arundinacea as comprising two polymorphic species, one cultivated (Sorghum bicolor) and the other wild (Sorghum arundinaceum). In Africa, however, the obvious convenience of such an approach has its drawbacks because there is continuous introgression between the two species, and the desirability of having a third, hybrid taxon (the oldest applicable name being Sorghum drummondii) is apparent. While Fl hybrids are clearly Intermediate, F2’s begin to show segregation and the boundry between the wild and cultivated Sorghums easily becomes heavily obscured.

It might be thought at first that such a problem would not arise in Asia because, according to Doggett (Sorghum, 1970), grain Sorghums were brought to India from Africa but the wild ones were not. The puzzle in Asia, therefore, is the occurrence of plants that are quite clearly forms of Sorghum arundinaceum; two such plants, Sorghum stapfii and Sorghum pugionifolium appear to exist only as type sheets. There are a number of possible explanations for their occurrence in India; one is that they are indeed endemic species (which seems unlikely) and another is that they are relatively recent introductions from Africa. A third possibility, perhaps the most favour-able, is that these two species, along with a handful of other collections of more typical Sorghum arundinaceum from India and Pakistan, are genetic throwbacks from the cultivated Sorghum bicolor. If they are, and this does seem possible, then there is a small nomenclatural problem to be overcome.

According to some observations of Doggett, hybrids between cultivated and wild Sorghums are unlikely to survive much beyond the third generation. The Fl intermediate plants would grow in fringe areas such as abandoned farmland, waste ground and roadsides, but the F2 plants would be subject to some kind of selection. In cultivated fields they would immediately be recognised as weeds and most likely uprooted and destroyed. In wild communities, however, plants with cultivated characteristics (such as persistent grains) will be at a selective disadvantage and such characteristics would probably be eliminated in one or two generations. This kind of evolutionary reversal would proceed to the extent that hybrid plants would eventually cease to be recognisable as such and best regarded as ordinary Sorghum arundinaceum. The name Sorghum drummondii should therefore be reserved for those transient uncultivated plants that display certain characteristics of Sorghum bicolor but which have arisen spontaneously. Selection of the kind outlined above would likewise operate on genetic throwbacks, restoring them eventually to their ancestral form, but plants of the first generation or two resembling an Fl hybrid cannot be morphologically distinguished from such hybrids and must therefore be called Sorghum drummondii. Certain segregates of Sorghum drummondii are cultivated for fodder and are known as Sudan Grass (Sorghum sudanense).

In China and parts of eastern India, Sorghum bicolor hybridises with the native diploid Halepensia, Sorghum propinquum, and the Kaoling Sorghums of China are thought to be segregates of this cross.

The spontaneous hybrid Sorghum halepense (Johnson Grass) x Sorghum bicolor is known as Sorghum almum Parodi. It is likely to occur in Asia but is apparently not cultivated for fodder here as it is in South America.