In this paper, Jennings studies chemotaxis in more detail. By “chemotaxis”, it is simply meant that paramecia tend to gather in certain substances and avoid others. This occurs mostly through the avoiding reaction. When Paramecium encounters an alkaline solution such as sodium chloride, it gives the avoiding reaction: it swims backward then turns, and swims forward again. This makes alkaline solutions repellent. The avoiding reaction also occurs when Paramecium leaves some solutions such as weak acids. When it enters a drop of acid, its course is unchanged; but when it reaches the border of the drop, it gives the avoiding reaction and therefore remains in the drop. Thus, chemotactic behavior is not the result of turning directly towards the stimulus, but rather of trial and error - changing direction when the chemical environment is undesirable.

First, Jennings examines the hypothesis that the avoiding reaction might be triggered by the toxicity of the substance, as in an “organic analogue of pain”. He demonstrates clearly that there is no general relation between toxicity and repelling power. He classifies repellents in two groups. Paramecia are repelled by very small concentrations of substances of the first group, well below their toxicity level. Substances of the second group only cause paramecia to react at toxic levels, when it is already too late (they die). For example, cane sugar induces plasmolysis, which then triggers the avoiding reaction, but the cell is already fatally injured. Jennings speculates that chemosensitivity to some repelling substances has been evolutionarily selected not because of their toxicity but because bacteria (Paramecium’s food) are typically not found in them.

The study illustrates some of the subtleties of examining chemotaxis. Chemosensitive behavior is not necessarily the result of sensing. For example, after some time, cane sugar induces plasmolysis, and then Paramecium begins to swim backward and forward repeatedly. Knowing the electrophysiology, we may hypothesize that plasmolysis depolarizes the cell, triggering repeated action potentials. Jennings describes Paramecium behavior mainly in inorganic solutions, but it is now known that the membrane is permeable to many of the small cations, and in addition calcium is the second messenger involved in the motor response. Therefore, some of the chemosensitive behavior observed by Jennings results from an interference between the substances and the normal function of the action potential. Some later work (in particular by Judith van Houten and coll.) has shown that Paramecium is also sensitive to a variety of large molecules, and appears to have chemoreceptors, for example to folate.