3.3. Track 2[[Student version, December 8, 2002]] 95
T 2 Track 2
3.3.2′ Sturtevant’s genetic map (Figure 3.13) also has a more subtle, and remarkable, property.
Choosing any three traitsA,B,andCappearing in the map in that order on the same linkage
group, the probabilityPACthatAandCwill be separated in a single meiosis is less than or equal
to the sumPAB+PBCof the corresponding probabilities of separation ofABandBC.There was
initially some confusion on this point. Requiring thatPACbeequaltoPAB+PBCled W. Castle to
propose a three-dimensional arrangement of the fly genes. Muller later pointed out that requiring
strict equality amounted to neglecting the possibility ofdoublecrossing-over. Revising his model to
incorporate this effect, and including later data, Castle soon found that the data actually required
that the genes be linearly arranged, as Morgan and Sturtevant had assumed all along!