<p>In the Appendix to the Transcendental Dialectic, Kant introduces several ideas of reason, e.g. those of pure air, pure water, and pure earth. I call these the “particular chemical ideas” in contrast to the three general transcendental ideas, namely, ‘God’, ‘world’, and ‘soul’. Among the few Kant scholars who study them, one predominant view is that we can discover instances of them, e.g., under laboratory conditions. This view denies that particular chemical ideas contain a unity that is permanently beyond the bounds of possible experience, which Kant introduces as a feature of ideas in general. According to this view, particular chemical ideas figure in scientific hypotheses that can be proved. When these hypotheses are proved, the ideas that figure in them are instantiated. By contrast, I argue that the particular chemical ideas do not figure in scientific hypotheses that can be proved. Instead, they are generated by what Kant calls the “hypothetical use of reason”, which is for him not how we produce provable hypotheses, but rather how we do metaphysics. The resulting particular chemical ideas, I argue, are best understood as denoting natural kinds—universals that are permanently beyond the bounds of possible experience.</p>