keeping the vowels free, so that they might be inserted between the consonants to form well-known words. His alphabet was: 1 = t; 2 = n; 3 = m; 4 = r; 5 = 1; 6 = d; 7 = c, k, g, q; 8 = b, h, v, w; 9 = p, f; o = s, x, z. From these equivalents the number 812 (I take it from the date of publication of his work, as a random example) could be represented by words such as "button," "obtain," or "Wotan."

Other teachers of memory systems—notably Aime Paris, Francis Fauvel Gouraud, Dr. Edward Pick, and others more recent, worked further upon this idea of number equivalents, introducing small improvements—mostly attempts to provide for each number a more or less equal representation. The lower case of a practical printer shows that certain letters are used in the English language much more frequently than others. Those which are comparatively little used should therefore be grouped in lots, each lot to represent one number.

I have studied most of these systems, and as a result have formed my own, which I believe to be a slight improvement upon even the best of any of the others. It happened that nearly twenty-five years ago I had a long illness, and during convalescence I had to lie down quietly for about six weeks. I took the opportunity during that time to study the combinations of the letters in all the commonly used words in the English dictionary.

Before I explain the method, in which I naturally adopted all that was best in the old systems, I must mention that the " fertile secret" was known among the Hindus long ago. I have before me a set of number-equivalents for the Sanskrit language given in Nilakantha's Commentary on the " Mahab-harata" (Adi Parva, end of Sarga 2). His system was called " Katapayadi." His number-equivalents were for consonants only as shown in the table on page 113.

I insert this only as a curiosity for European readers, and

0 0

Post a comment