:: :: University of Virginia Library

to be made clear. As their format has been discussed in some detail elsewhere, however,[7] a summary will suffice. In order, the manuscripts are (1) the 'Classified Papers' (hereinafter called CP), the first drafts, often holograph, of scientific matters read or communicated to the Society, of which most were transcribed by the Secretaries into (2) the Register (R), which consists of several folio volumes. The other manuscript source (we disregard the Society's Letter-Book) is interesting as regards its provenance and independent value: it consists of transcripts of selected materials from the CP or from the Register made by copyists for Sir Hans Sloane and these now form certain of (3) the Sloane MSS in the British Museum (Slo.). It is not always clear what their copytext was; in some instances it seems no longer extant, and the reasons for some of their variants, considered later, are often indeterminate: some seem to be honest endeavours to correct or to clarify a corrupt text, as happened in the later Shakespearan Folios, were these manuscript; others are clearly no more than errors in transcription.

Our printed sources are, firstly, those pages of Thomas Sprat's History of the Royal Society which constitute versions of selected scientific materials contributed by Fellows and form its Second Part. The four editions (1667, 1702, 1722, 1734; the symbol HRS below should be taken as referring to the first, which alone is of any authority) are for practical purposes identically paged. A facsimile reprint of this first edition[8] records only significant variants in the scientific papers; it does not discuss the reasons for any corruptions. In general, variants there available for examination are not repeated here. Next we have Thomas Birch's four volumes (1753-56) of the same title. It suffices our present purpose to note that this work, in the main, consists of transcripts from the Society's Journal illustrated and supplemented by extracts from the Register and from the Letter-Book. What formed the copy for the relevant parts of Sprat and for Birch is unknown: with the former it was probably a fair copy made by clerks either from the CP or from R; Birch probably prepared his own from the same sources, but there is some evidence, in the form of alterations in his hand in R, that he used Sprat where he could. Accordingly he is used little here. Apart from the occasional instance in the later editions of Sprat, noted below, no instance has been found in connection with the

present enquiry where any printed text is of independent authority — one trivial instance may be ignored.[9] The last printed source is the Society's Philosophical Transactions which, broadly speaking, consist of edited versions of scientific papers, much as in Sprat, but published regularly. Most changes are clearly deliberate, and frequently go as far as a complete re-casting of materials. As a couple of small examples we may take the 'Answers return'd by Sir Philiberto Vernatti, Resident in Batavia in Java Major, to certain enquiries. . . .' (Sprat, 158ff.), of which certain of the questions figure in the Transactions (11 March 1666/7, 415-419) in much altered form, and the 'Direction for the observations of the eclipses of the moon', of which the version in Sprat (180 ff.) and in the Transactions (11 February 1666/7, 388-390) differ only in occasional alteration of word-order and other small details.

Fortunately in the printed materials to be dealt with there appear to be no examples of the 'correction-of-a-correction' error (discussed fully by McKerrow in the unpublished lectures already noted) which, if I may recall his argument, are errors at two removes: the compositor cannot read his copy and conjectures, or else he accidentally mis-sets, and a second person, often the proof-reader, sensing an error in the printed proof, alters it yet again, but from his own judgment and without consulting the copy, so causing a double corruption and one very hard to correct once the true manuscript copy has been lost. Finally it must be made clear that only a small sample from the vast manuscript resources available is here being considered — those papers which happen to occur in Birch, Sprat and the Transactions, or in some of them. Few general conclusions will be attempted at the end.