398 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1937 



the panicle. Although the idea of branch coherence was proposed 

 for the ear it would account equally well for the central spike of the 

 maize tassel. This theory is supported by the recurrence of bifurcated 

 and fasciated ears and central spikes but has obtained no histological 

 confirmation (pi. 21). 



A second suggestion has been made that the central spike came 

 about through the suppression or reduction of tassel branches into a 

 simple pair of spikelets, thus leaving the central axis surrounded by 

 paired spikelets. This theory derives plausible support from certain 

 peculiar branched forms of maize where the branches extend in an 

 unbroken series of ever-decreasing lengths from the base to the tip 

 of the tassel. In this form of maize, known as ramose, and in certain 

 other similar types, such as branched silkless, all stages between 

 true branches and paired spikelets are found. Where these forms 

 have been hybridized with normal maize there often is, in the genera- 

 tions subsequent to the first, a great variety of intermediate forms 

 showing all degrees of transition from tassel branches to paired 

 spikelets (pi. 22). 



Then, too, among the podded forms of maize it is not uncommon 

 for a pair of spikelets to develop into a branch. The process is also 

 seen on both ears and tassels in a variation where the flowers are 

 proUf crated. 



A third method of deriving the ear has been suggested from a 

 study of the hybrids between com and teosinte where it has been 

 shown that a yoking of spikelets takes place. In some plants these 

 yokes fit together at right angles, thus constituting an eight-rowed 

 ear (pi. 23), while in others the yokes occur in more complicated 

 figures, giving ears with higher numbers of rows. Tliis theory, 

 while necessitated by the behavior of com-teosinte hybrids, does 

 not easily account for ears with rows of kernels in multiples of two. 

 This same objection has been raised to the theory of branch coalescence 

 because each branch is a 4-rowed unit and the union of two such 

 branches would give an 8-rowed ear; three would give a 12-rowed 

 ear, etc. The objection, based on the undoubted fact that ears 

 having 10, 14, 18, and 22 rows are common, is pertinent but not 

 insurmountable, for the commonest feature of corn and its relatives 

 is one of abortion or suppression of parts. Leaf blades, branches, 

 spikelets, and flowers are regularly aborted or rather suppressed, 

 and the suppression of a single row of paired spikelets is sufficient to 

 reduce a 12- to a 10-rowed ear. This suppression of rows of paired 

 spikelets is a commonplace in most cornfields. Indeed, ears that 

 begin with eight rows often through poor growing conditions end as 

 six-rowed ears and even as four-rowed ones (pi. 24). The phenomenon 

 is common in ears having higher numbers of kernel rows. 



