The most detailed and diverse data available from classical and especially allelic crossing over are examined on the basis of a hybrid DNA model of crossing over. The model is similar to that proposed byHolliday. There are three main points of difference: (1) It is proposed that crossing over begins with a switch of chromatids, as in the classical hypothesis. Dissociation of DNA then begins from the nearest recombinator and extends as far as the switch. The switch, therefore, limits the spread of hybrid DNA formation. (2) Since dissociation begins from the nearest recombinator, the midpoint between two recombinators is the point about which the probability of dissociation, either from the left or the right, is determined. (3) The probability of breaks in the chains that dissociated is higher near the recombinators, and the probability of breaks of the chains that did not dissociate is higher near the mid-point. The overall ratio of the two types of break is, however, 50∶50. The observations the model attempts to explain include the relationship between chiasmata and crossing over; positive interference; negative interference; polarity; differential rates of conversion.