Frequencies of bigrams (two-letter combinations) in written language, e.g., Underwood and Schulz (1960), may be used to compute total bigram frequency of any set of letters by summing, left to right, frequencies of successive bigrams. Mayzner and Tresselt (1959) showed that, if the same word was rearranged into two anagrams, one with high, the other wich low, total bigram frequency, solution times were significantly slower for anagrams with high totals. They suggested that associative is greater between letters of high frequency bigrams, so it is more difficult to break up such pairs when actempting to rearrange letters of an anagram. It is also possible that cohesive strength between adjacent letters may in part depend on physical proximity of the letters. If so, anagrams with wide spacing between individual letters may be more easily solved than anagrams with letters close together; this hypothesis is tested here. The five, Very Frequent (Thorndike-Lorge count) words, chair, sugar, train, party, and labor, and the five Very Infrequent words, tango, groin, peony, ghozll, and triad, used by Mayzner and Tresselt in another study (1958), were employed. After rearrangement into anagrams, each anagram was typed in capital letters on two cards. On one card the space between successive letters was that produced normally by an elite-type typewriter (Nonspaced Condition). On the other card successive letters were separated by 10 typewriter spaces (Spaced Condition). In each condition 36 undergraduate Ss attempted to solve all 10 anagrams. Two letter orders were used for each anagram. The two orders selected were the one that yielded the highest, and the one that yielded the lowest, bigram frequency totals obtainable within the five hard letter orders developed by Mayzner and Tresselt (1958). Within each condition, half the Ss were given the version with high total bigram frequency, half the version wich low total frequency, of each of the 10 anagrams. E showed S a sample anagram, then presented each of the 10 anagrams one at a time, using a different random order of presentation for each S. S gave his solutions verbally. If S had not solved an anagram in 4 min., E told him the solution and presented the next anagram. The time of successful solutions was recorded in seconds. Since there were a number of failures to solve in 4 rnin., the data analyzed