86 Chapter 3. The molecular dance[[Student version, December 8, 2002]]
time.) But whatwerechromosomes, anyway? It seemed impossible to make further progress on
this point with the existing cell-biological tools.
Asurprising quirk of genetics broke the impasse. Though Mendel’s rules were approximately
correct, later work showed that not all traits assorted independently. Instead, W. Bateson and
Correns began to notice that certain pairs of traits seemed to belinked,aphenomenon already
predicted by Sutton. That is, such pairs of traits will almost always be inherited together: The
offspring gets either both, or neither. This complication must have seemed at first to be a blemish
on Mendel’s simple, beautiful rules. Eventually, however, the phenomenon of linkage opened up a
new window on the old question of the identity of genetic factors.
The embryologist T. H. Morgan studied the phenomenon of genetic linkage in a series of exper-
iments starting around 1909. Morgan’s first insight was that in order to generate and analyze huge
sets of genealogical data, big enough to find subtle statistical patterns, he would need to choose a
very rapidly multiplying organism for his model system. Certainly bacteria multiply rapidly, but
they were hard to manipulate individually and lacked readily identifiable hereditary traits. In a
fateful decision, Morgan’s compromise choice was the fruit flyDrosophila melanogaster.
One of Morgan’s first discoveries was that some heritable traits in fruit flies (for example, white
eyes) were linked to the fly’s sex. Since sex was already known to be related to a gross, obvious
chromosomal feature (females have two X chromosomes while males have just one), the linkage of
amutable factor to sex lent direct support to Sutton’s and Boveri’s idea that chromosomes were
the physical carriers of Mendel’s factors.
But now an even more subtle level of structure in the genetic data was beginning to appear. Two
linked traitsalmostalways assorted together, but would occasionally separate. For example, certain
body-color and eye-color factors separate in only about 9% of offspring. The rare failure of linkage
reminded Morgan that F. Janssens had recently observed chromosome pairs wrapping around each
other prior to meiosis, and had proposed that this interaction could involve the breakage and
exchange of chromosome pieces. Morgan suggested that thiscrossing-overprocess could explain
his observation of incomplete genetic linkage (Figure 3.12). If the carrier object were threadlike,
as the chromosomes appeared to be under the microscope, then the genetic factors might be in
afixed sequence, or linear arrangement, along it, like a pattern of knots in a long rope. Some
unknown mechanism could bring two corresponding chromosomes together and align them so that
each factor was physically next to its partner, then choose a random point at which to break and
exchange the two strands. It seemed reasonable to suppose that the chance of two factors on the
same chromosome being separated by a physical break should increase, the more distant their fixed
positions were. After all, when you cut a deck of cards the chance of two given cards getting
separated is greater, the farther apart they were to begin with in the deck.
Morgan and his undergraduate research student A. Sturtevant set themselves the task of ana-
lyzing these exceptions. They reasoned that if the hypothesis of linear sequence were true, then
for each set of linked traits it should be possible to list those traits along a line in such a way that
the probability of two traits’ getting separated in an offspring was related to their distance on the
line. Examining the available data, Sturtevant found that not only was this the case, but moreover
each linked set of traits admitted just one such linear arrangement fitting the data (Figure 3.13).
Two years later the dataset had expanded to include 24 different traits, which fell intoexactly four
unlinked groups—the same number as the number of visible chromosome pairs (Figure 3.13)! Now
one could hardly doubt that chromosomes were the physical objects carrying genetic factors. The
part of a chromosome carrying one factor, the basic unit of heredity, was christened the “gene.”