Mr. Bartley's remark that Goodman's puzzle ([2], Chap 3) is “an interesting variant of the possibility .. . that the next instance may be different” ([1], p. 85) rather badly misrepresents matters. One might say that the “new riddle” arises just because no matter what the nature of the next instance, it will be as much like all previously examined cases as any other instance. Suppose “Hester” is the name of the first emerald examined after time t. If Hester is green, then Hester is like previously examined emeralds with respect to their green-ness. But should Hester be blue, then Hester has its grue-ness in common with previously examined emeralds. The difficulty here is that likenesses abound; the evidence presents us with more likenesses than we could consistently expect of or project to subsequent cases. And the question that Goodman's puzzle poses is, by what criteria do we select one from among all of these likenesses to project to subsequent cases ?