arrangement of the characters in the De Finibus is announced later still[152]; and even at a later date Cicero complains that Balbus had managed to obtain surreptitiously a copy of the fifth book before it was properly corrected, the irrepressible Caerellia having copied the whole five books while in that state[153].  A passage in the De Divinatione[154] affords almost direct evidence that the Academica was published before the De Finibus.  On all these grounds I hold that these two works cannot be those which Cicero describes as having been finished simultaneously at Astura.

Another view of the [Greek:  syntagmata] in question is that they are simply the two books, entitled Catulus and Lucullus, of the Priora Academica.  In my opinion the word [Greek:  syntagma], the use of which to denote a portion of a work Madvig suspects[155], thus obtains its natural meaning.  Cicero uses the word [Greek:  syntaxis] of the whole work[156], while [Greek:  syntagma][157], and [Greek:  syngramma][158], designate definite portions or divisions of a work.  I should be quite content, then, to refer the words of Cicero to the Catulus and Lucullus.  Krische, however, without giving reasons, decides that this view is unsatisfactory, and prefers to hold that the Hortensius (or de Philosophia) and the Priora Academica are the compositions in question.  If this conjecture is correct, we have in the disputed passage the only reference to the Hortensius which is to be found in the letters of Cicero.  We are quite certain that the book was written at Astura, and published before the Academica.  This would be clear from the mention in the Academica Posteriora alone[159], but the words of Cicero in the De Finibus[160] place it beyond all doubt, showing as they do that the Hortensius had been published a sufficiently long time before the De Finibus, to have become known to a tolerably large circle of readers.  Further, in the Tusculan Disputations and the De Divinatione[161] the Hortensius and the Academica are mentioned together in such a way as to show that the former was finished and given to the world before the latter.  Nothing therefore stands in the way of Krische’s conjecture, except the doubt I have expressed as to the use of the word [Greek:  syntagma], which equally affects the old view maintained by Madvig.

Whatever be the truth on this point, it cannot be disputed that the Hortensius and the Academica must have been more closely connected, in style and tone, than any two works of Cicero, excepting perhaps the Academica and the De Finibus.  The interlocutors in the Hortensius were exactly the same as in the Academica Priora, for the introduction of Balbus into some editions of the fragments of the Hortensius is an error[162].  The discussion in the Academica Priora is carried on at Hortensius’ villa near Bauli; in the Hortensius at the villa of Lucullus near Cumae.  It is rather surprising that under these circumstances there should be but one direct reference to the Hortensius in the Lucullus[163].