1. Omnis homo, equus, asinus, &c., est longævus. 2. Omnis homo, equus, asinus, &c., vacat bile. Ergo: 3. Quicquid vacat bile, est longævum. Convertible into a Syllogism in Barbara:-- 1. Omnis homo, equus, asinus, &c., est longævus. 2. Quicquid vacat bile, est homo, equus, asinus, &c. Ergo: 3. Quicquid vacat bile, est longævum. Here the force of the proof (or the possibility, in this exceptional case, of converting a syllogism in the Third figure into another in _Barbara_ of the First figure) depends upon the equation or co-extensiveness (not enunciated in the premisses, but assumed in addition to the premisses) of the minor term C with the middle term B. But I contend that this is _not_ the condition peremptorily required, or sufficient for proof, if we suppose C the minor term to represent _omne longævum_. We must understand C the minor term to represent _omne vacans bile_, or _quicquid vacat bile_: and unless we understand this, the proof fails. In other words, _homo, equus, asinus, &c._ (the aggregate of individuals), must be co-extensive with the class-term bile-less or _vacans bile_: but they need not be co-extensive with the class-term long-lived or _longævum_. In the final conclusion, the subject _vacans bile_ is distributed; but the predicate _longævum_ is not distributed; this latter may include, besides all bile-less animals, any number of other animals, without impeachment of the syllogistic proof. Such being the case, I think that there is a mistake in the text as given by all the editors, from Pacius down to Bekker and Waitz. What they give, in setting out the terms of the Aristotelian Syllogism from Induction, is: [Greek: e)/stô to\ A makro/bion, to\ d' e)ph' ô(=| B, to\ cholên mê\ e)/chon, e)ph' ô(=| de\ G, _to\ kath' e(/kaston makro/bion_, oi(=on a)/nthrôpos kai\ i(/ppos kai\ ê(mi/onos.] Instead of which the text ought to run, [Greek: e)ph' ô(=| de\ G, _to\ kath' e(/kaston a)/cholon_, oi(=on a)/nthr. k. i(/p. k. ê(mi/]. That these last words were the original text, is seen by the words immediately following: [Greek: tô=| dê\ G o(/lô| u(pa/rchei to\ A. _pa=n ga\r to\ a)/cholon makro/bion_]. For the reason thus assigned (in the particle [Greek: ga/r]) is irrelevant and unmeaning if [Greek: G] designates [Greek: to\ kath' e(/kaston _makro/bion_ ], while it is pertinent and even indispensable if [Greek: G] designates [Greek: to\ kath' e(/kaston _a)/cholon_]. Pacius (or those whose guidance he followed in his text) appears to have perceived the incongruity of the reason conveyed in the words [Greek: pa=n ga\r to\ a)/cholon makro/bion]; for he gives, instead of these words, [Greek: pa=n ga\r _to\ G_ makro/bion]. In this version the reason is indeed no longer incongruous, but simply useless and unnecessary; for when we are told that A designates the class _longævum_, and that [Greek: G] designates the individual _longæva_, we surely require no reason