The invention was not altogether a success, very heavy bags—especially when the trains were running at great speed—being sometimes held responsible for the occurrence of rather serious accidents. It even became necessary to cease using the apparatus till the defect, whatever it might be, could be put right. Several remedies were suggested, but none proved effectual till my brother, then only twenty-one years of age, hit upon a simple contrivance which removed all difficulties, and thenceforth the exchange-bag apparatus worked well. Sir William Cubitt, who had unsuccessfully striven to rectify matters, generously eulogised his youthful rival's work. The stamp-obliterating machines which superseded the old practice of obliteration by hand were also my brother's invention. In former days the man who could stamp the greatest number of letters in a given time was usually invited to exhibit his prowess when visitors were shown over the office. The old process had never turned out impressions conspicuous for legibility, and means of improvement had been for some time under consideration. But it was a trial presided over by Lord Campbell in 1856 which precipitated matters. An important question turned upon the exact date at which a letter had been posted, but the obliterating stamp on the envelope was too indistinct to furnish the necessary evidence. Lord Campbell sharply animadverted upon the failure, and his strictures caused the Duke of Argyll—then Postmaster-General—to write to Rowland Hill upon the subject. The use of inferior ink was supposed to be responsible for the trouble, and various experiments were tried, without effecting any marked beneficial result. Objection was made to abolition of the human hand as stamper on the ground that thus far it had proved to be the fastest worker. Then my brother's mechanical skill came to the rescue, and complaints as to clearness and legibility soon became rare.[198] By the machines the obliterations were made faster than by the best hand-work, the increase of speed being at least 50 per cent. About the year 1903 my brother's machines began, I am told, to be superseded by others which are said to do the work faster even than his. Judging by some of the obliterations lately made, presumably by these later machines, it is evident that, so far as clearness and legibility are concerned, the newer process is not superior to the older.