WHAT IS A GENE? 167 



8 X 10^ nucleotide pairs. There seems to be just one linkage group 

 in phage T4 with a length of about 800 map units, corrected for suc- 

 cessive rounds of mating and for interference. Dividing 800 map units 

 by 8 X 10^ nucleotide pairs gives 0.01 map units per nucleotide pair. 

 That is to say, if two mutants, having mutational changes one nucleotide 

 apart, are crossed, the proportion of recombinants in the progeny should 

 be 0.01 per cent. 



This is a very exciting result for the maximum sizes of the recon and 

 of the muton approach very closely the length of one nucleotide pair. 

 We should recall the evidence on induction of mutation in TMV by 

 nitrous acid, where a change in one nitrogen base seemed capable of 

 causing a mutation. Nonetheless, there are a fair number of assumptions 

 in the calculation of muton and recon size. It was gratuitously assumed, 

 among other things, that the probability of recombination per unit 

 molecular length was constant and that the vagaries of interference 

 were appropriately assessed. Without doubt, as our knowledge becomes 

 more refined, such calculations will be improved, but it now seems prob- 

 able that revision will be in the direction of establishing the muton as 

 a single nucleotide pair. This is what we considered it to be in the dis- 

 cussion of mutation (Chapter 2). Recombination presumably occurs be- 

 tween sequential couplets of nucleotide pairs. But whether it does so 

 by actual breakage and reformation of the sugar-phosphate backbone of 

 DNA or by some replication mechanism is another question. 



A CONCEPT OF THE GENE 



The functional unit is certainly the complex one. On the basis of cal- 

 culations like those made above, it should consist of hundreds to 

 thousands of nucleotide pairs. Its identification by the cis-trans test 

 allows the sharpest definition of its limits, even when the enzymology of 

 its effect is unknown. As a consequence, it is named the cistron. But 

 even the test of complementation is not absolute, and results of the cis- 

 trans test must always be scrutinized critically. Cases are known where 

 mutons within the same cistron partially complement one another. One 

 instance is found in Aspergillus where, at the ad^ locus on chromosome I, 

 there is a cistron containing six identified mutational sites. Some of these 



are noncomplementary, e.g., ^— "^ — ^. But the fraas-configuration 



+ a«i5 



^^-^ — ^ grows a little on minimal medium, showing a partial comple- 



+ ««15 



mentation. In another case, two Neurospora mutants aff^ecting the 



