Discusses P. S. Pallas’ theory of origin of domestic dog breeds. CD believes domestic dogs descended from more than one aboriginal wild species but ultimately "we believe all canine species have descended from one parent and the only question is whether the whole or only part of difference in our domestic breeds has arisen since man domesticated them".

Races of man offer great difficulty. The doctrine of Pallas and Agassiz that there are several species "does not help us" in the least.

Hopes Henry Holland will not review Origin.

CD’s and CL’s difference on "principle of improvement" and "power of adaptation" is profound. Improvement in breeds of cattle requires neither. Urges him to reread first four chapters of Origin carefully. Natural selection is not to be contrasted with "improvement": every step involves improvement in relation to the conditions of life. There is no need for a "principle" to intervene.