POLYNUCLliOTlDU SYNTIIUSIS IN NUCLIiOLUS AND CI IKOMOSOM l.S 1 73 



poration, labeled nuclei appear at (li\isi()ii stages. The chromosomes in these first 

 labeled cells which appear at division are sometimes nonuiiitormly labeled along 

 their length; sometimes parts of chromosomes, and very rarely whole chromo- 

 somes may be unlabeled in an otherwise labeled group (12; Woods, unpublished). 

 This observation suggests that duplication does not begin simultaneously at all 

 points along a chromosome and, further, that variations may occur among indi- 

 vidual chromosomes of a set. Cells appearing at divisions a few hours later have 

 their chromosomes much more uniformly labeled. The first labeled cells to reach 

 division have evidently completed part of DNA synthesis before coming in con- 

 tact with the isotope, but after a few more hours only those which have been in 

 contact with the isotope for the entire synthetic period are arriving at division 

 stages. There appear to be variations among species in the degree aiitl pattern 

 of asynchrony. In Crcpis, a plant with three pairs of relatively small chromosomes, 

 the duplication appears to begin at both ends of a chromosome and progresses 

 toward the centromere in a zipper fashion (12). In Bellevalia, a plant of the lily 

 family with larger chromosomes, the amount of asynchrony is small. Nearly all 

 chromosomes are labeled simultaneously. No exact measure of the time required 

 for duplication is available, but in Bellevalia the period is a matter of from 4 

 to 8 hours at 25 °C. This includes one-third or less of the interval between 

 division stages. 



At the first division after labeled thymidine has been incorporated, each of 

 the two daughter chromosomes is similarly labeled and the amount of label in 

 each can be shown to be equal within the limits of error of the method (13). 

 Figure I shows a cell at the first division after treatment with colchicine to 

 prevent spindle formation. The daughter pairs still lie close together but have 

 separated in most instances. When the labeled thymidine is removed and the 

 labeled chromosomes allowed to pass through a second duplication, they regularly 

 yield one labeled and one non-labeled daughter chromosome (fig. 2). The pool 

 of labeled precursors fortunately is small and is depleted before the second 

 duplication. 



The observed behavior is adequately explained if the original chromosomes are 

 assumed to be composed of two units of DNA. When duplication occurs in 

 labeled thymidine each daughter receives one original unit and one new labeled 

 unit, and, therefore, both appear labeled. However, when these labeled chromo- 

 somes duplicate without the label, the two units separate and along each is built 

 an unlabeled unit. Since only one unit of each chromosome is labeled they pro- 

 duce one labeled daughter and one unlabeled daughter. The apparent exceptions 

 (fig. 2) are those in which exchange of segments occurs between daughter 

 chromosomes (chromatids) before they appear at division. Since this is a rather 

 frequent event, chromosomes usually appear labeled along part of their length 

 and unlabeled along the remainder. However, when account of the exchanges 



