Obviously, the easiest way to manufacture this error is to select a stamp from plate 2 with the lettering of "L.F.--F.L." (the last stamp in the _sixth_ row), and alter the first "F" into "H", with hope of probable success because the collector's criticism would naturally (if wrongly) be concentrated on the incorrect letter in the lower left-hand corner. Unfortunately for the "fake," which was very well executed, its creator, wishing no doubt to enhance its value, had left the "error" in pair with the eleventh stamp in the same row: result, a very nice pair from the sixth row, lettered "K.F.--F.K.", "L.H.--F.L.", showing (as a consequence of being in pair) a mistake--"H" for "F" in the upper right-hand corner. This, of course, condemned the error at once, but the example serves to show how very careful one must be, and how necessary it is to examine and consider every circumstance in connection with the particular stamp under observation. There are two varieties of stamps, differing from the normal through some slip in the process of manufacture--bicoloured stamps, in which the portion printed in one colour is inverted as regards the remainder of the design, caused by carelessness in "feeding" the partly-printed sheet wrong way up into the press, for the second impression completing the design; and pairs of stamps, which, each quite normal if severed, are when _se tenant_ inverted in respect to each other, a condition philatelically termed _tête-bêche_. The fraudulent manipulator has turned his attention to these, generally scarce and frequently very rare, eccentricities, cutting out from the bicoloured stamp the part printed in one colour and replacing it with great care, but upside down; and, as to the _tête-bêche_ pairs, manufacturing them by means of two single copies, a strong adhesive mixture and heavy pressure. Sometimes, so well have these frauds been made that nothing short of several hours' _boiling_ has sufficed to dissolve the illegal union of the two pieces of paper--a drastic test, and one somewhat detrimental to the value of such copies as are enabled, by their genuineness, to survive the ordeal. The possible result to, say, a mint imperforate Fourpence, Ceylon, suspected of having recently acquired its otherwise desirable "margins," reminds me of the test given (not advocated) by a famous philatelist for the detection of forgeries of early Cashmere stamps, which were printed in water-colour--"Put them in water; if the colour is 'fast' the stamp is a forgery; if it comes off, leaving a blank piece of paper, the stamp is genuine"!