termediate, though more like the normal parent, and practically 
just as able to care for themselves. There were no shepherd’s 
crook or ram’s horn distortions in these plants and they were as 
like each other as beans in a pod, though easily distinguished 
from either parent. But when offspring resulting from their 
guarded seed were raised, only approximately one half resembled 
the parent, while approximately one-fourth looked like the grand- 
mother and one-fourth like the grandfather. Whenever guarded 
seed (seed from bagged flowers) was sown from either of these 
two latter, they always produced plants that one could not dis- 
tinguish from their respective parents. However, seed from those 
grandchildren characterized as intermediate, always produced 
the three kinds of plants just mentioned, in approximately the 
proportions 1:2:1. From seven of these intermediates, 465 off- 
spring were grown, 115 of which were similar to the original ab- 
normal Cuban ancestor, 228 were intermediates, and 122 were 
normal plants, in all respects similar to the ancestor of the 1907 
mutant plant. From experiments extending over five generations, 
we have found that these normal and abnormal plants breed true 
from generation to generation just as though they had never come 
from a mixed parentage. On the other hand, the intermediates 
of each generation have always thrown offspring of the three kinds 
and in the proportions of 1:2:1. From all we know at present 
about the inheritance of characters in plants and animals, this 
production of the three kinds by the intermediates would go on 
indefinitely, and the two small classes would breed true indefinitely, 
barring mutations and changes due to environment. From the 
facts just presented regarding what transpires when tvro such 
true-breeding plant types are crossed, we feel justified in asser- 
ting that this mutant race differs from the race that gave it birth 
in a single hereditary unit (or character differences that are in- 
herited as a single unit) which is capable of expressing itself as 
a large number of taxonomic characters. For as we have seen, 
the distinguishing features of this strain are inherited as though 
they were all tied together. No matter how often these experi- 
ments wese repeated with these strains, the same results were 
always secured. After seeing this demonstrated anew from 
time to time, one comes to feel that we can be almost as certain 
of the result, as when one multiplies, adds, or subtracts the 
figures 2 and 4. This mating of the monstrous strain with indi- 
viduals of a normal strain similar to its own normal ancestors 
marks the second of the great events in the history of this family. 
The third great event resulted from matings of members of the 
monstrous strain with normal plants outside the family,— morg- 
anatic marriages, so to speak. The plant “classificationists” will 
tell you that all tobacco plants are normal that have characters 
