The stamps were printed in sheets of 240, divided into 4 panes each of 10 rows of 6 stamps and using two printing plates, one for the head of King George V, known as the head plate or vignette and a second for the frame, which was known as the duty plate. One duty plate was used for each value with the exception of the 25 cent value which was used until 1919 when it was discovered that it contained a wrong Chinese character and a new plate was made. One common head plate was used for all values and sheets exhibit just that number. Consequently the only plate number that appears on all of the China overprints is that of Head Plate Number 1 which appears on the panes above 1-2 NW and 1-5 NE and below 10-2 SW and 10-5 SE.
While attempts to combine requisitions for adhesives to be overprinted with printings of the regular Hong Kong issues were made, on a number of occasions the stamps to be overprinted were printed separately. As a result, some varieties and errors of printing can be found on both the overprinted and unoverprinted stamps but some are unique to the CHINA overprints.
Once the sheets of stamps had been printed by De La Rue, they were set aside by the Crown Agents' Stamp Examiners and then forwarded to Somerset House where the CHINA overprint and the Requisition Letter and Sheet Number were applied. (Perrin p14) After the stamps were overprinted they were returned to the GPO London who forwarded them to the Postmaster General in Hong Kong who sent them to the individual agencies.
From Perrin:
'Both Surgeon-Captain Bishop and Colonel Webb pointed out that some sheets of the normal Hong Kong series were divided before dispatch, in which case' the "requisition details may be found on left panes or in the 'stamps paper' margin of lower panes". In such cases, each half was composed of either the two upper of the two lower panes.
There is ample evidence that this practice also applied to the CHINA series, at least in the early printings. The Requisition A (1917) 25c. sheet used for the initial study of the overprint setting (see Plate) comprises in the N.W. and N.E. panes only, the bottom margin having been bisected and showing no sign of perforations. As regards the 1916 printing, Inland revenue records show that the consignment delivered to Somerset House for overprinting in many cases included not only some half-sheets but also quarter-sheets. A detailed breakdown is given in Appendix D.
The rationale of these partial sheets is by no means clear. In correspondence between the G.P.O. and Hong Kong there is no evidence that Hong Kong asked for them, and returns dealt only with full sheets of 240-though these must have included partial sheets, since the totals of '240' forwarded to Hong Kong in most cases exceeded the number of full sheets overprinted. The breakdown to half- and quarter-sheets is found only in correspondence between the G.P.O. and Somerset House. Moreover, the quantities themselves are perplexing. It might have been expected that the high and the little-used dollar denominations would be subjected to this bisection in order to facilitate distribution in smaller quantities; whereas these were only printed in full sheets. Nor is it clear why, among the lower values, the 8c. and 25c. alone escaped bisection (especially as the 25c was bisected in the 1917 printing); nor why, although the 2c. and 4c. denominations were produced in almost equal quantities, nearly twice the number of half- and quarter-sheets were printed for the 2c compared with the 4c. I can only present the available facts for speculation and further evidence.'