Instead of searching for the sense of each of the historian's words, and for the thought he has expressed in them, he [M. Monod] comments on each sentence in the light of what is found in Tacitus or the Salic law.... We should understand what analysis really is. Many talk about it, few use it.... The use of analysis is, by an attentive study of every detail, to elicit from a text all that is in it; not to introduce into the text what is not there." After reading this excellent advice it will be instructive to read M. Monod's reply (in the _Revue historique_); it will be seen that Fustel de Coulanges himself did not always practise the method he recommended. [133] Cf. _supra_, p. 103. [134] The work of analysis may be entrusted to a second person; this is what happens in the case of _regesta_ and catalogues of records; if the analysis has been correctly performed by the compiler of _regesta_, there is no need to do it over again. [135] Practical examples of this procedure will be found in Deloche, _La Trustis et l'antrustion royal_ (Paris, 1873, 8vo), and, above all, in Fustel de Coulanges. See especially the study of the words _marca_ (_Recherches sur quelques problêmes d'histoire_, pp. 322-56), _mallus_ (ibid., 372-402), _alleu_ (_L'Alleu et le domaine rural_, pp. 149-70), _portio_ (ibid., pp. 239-52). [136] The theory and an example of this procedure will be found in Fustel de Coulanges, _Recherches sur quelques problêmes d'histoire_ (pp. 189-289), with reference to the statements of Tacitus about the Germans. See especially pp. 263-89, the discussion of the celebrated passage on the German mode of culture. [137] Fustel de Coulanges formulates it thus: "It is never safe to separate two words from their context; this is just the way to mistake their meaning" (_Monarchie franque_, p. 228, note I). [138] This is how Fustel de Coulanges condemns this practice: "I am not speaking of pretenders to learning who quote second-hand, and at most take the trouble to verify whether the phrase they have seen quoted really occurs in the passage indicated. To verify quotations is one thing and to read texts quite another, and the two often lead to opposite results" (_Revue des questions historiques_, 1887, vol. i.). See also (_L'Alleu et le domaine rural_, pp. 171-98) the lesson given to M. Glasson on the theory of the community of land: forty-five quotations are studied in the light of their context, with the object of proving that none of them bears the meaning M. Glasson attributed to it. We may also compare the reply: Glasson, _Les Communaux et le domaine rural a l'époque franque_, Paris, 1890. [139] All that is original in Fustel de Coulanges rests on his interpretative criticism; he never did personally any work in external criticism, and his critical examination of authors' good faith and accuracy was hampered by a respect for the statements of ancient authors which amounted to credulity.