THE DISTINCTIVE QUALITY of children's books lies in their selectivity: in what has been left out, in deference to bodies still too small to pick up an unabridged dictionary. Such everyday adult words as "oxymoron" and "syzygy" have no place in these books, although one would immediately discard a college dictionary which lacked them.

There used to be considerable guesswork in deciding what to omit from a children's dictionary, and there seems to be some still when the dictionary is for beginning readers. But by the time they get to the third grade, sociology and computerization have tracked what can be expected of the average child. Thus The Children's Dictionary can claim that its 30,000-plus entries are "chosen on the basis of computer studies of books children actually read." There is no serious reason to doubt this claim, I suppose. I would like to think that there are still quite a few children who encounter such words as "bilge" and "forecastle" (or even "fo'c's'le") in their reading before they finish the ninth grade, and that there are still 12-year-olds who will feel the need to look up "towhead," "dragoon" and "extirpate." They will not find them in The Children's Dictionary, but they should be able to find them elsewhere by the time the need arises -- or, more likely, to figure out what they need to know without using a dictionary.

Such eccentricities aside, it seems likely that The Children's Dictionary includes the 30,000 words most likely to be needed most frequently by most children in grades three to nine. It is logically arranged and abundantly illustrated; the definitions are clear and concise, it is prefaced with an engaging little essay on the history of the English language, and each letter of the alphabet is introduced with a little essay on its ancestry in Phoenician, Greek, Roman and medieval writing. For some reason, the earlier forms of the letter "L" are omitted in this edition, and the historic treatment of such letters as "X" and "W" may be moderately misleading because space was lacking to give the whole story. But all in all, this is an expert, well-intentioned dictionary that is likely to give good service to the average child in elementary school. If it sometimes reads like the work of a computer rather than of a person who has loved our children's classics, and if it offers sad testimony on the decline of literacy in our time, that is presumably as it should be. Dictionaries are no longer expected to set or enshrine standards, but to tell us what is happening to the language. The picture in this dictionary is probably, on the whole, accurate.

At the entry level, in dictionaries for children and roughly three to eight, a simple comparison soon uncovers serious disagreements on basic policy. Under "monster," for example, My First Dictionary states baldly that "A monster is a huge, evil animal or person," while the folks at Sesame Street (who deal more regularly with monsters) take a more benign view: "A monster is a large or strange living thing. The monsters on Sesame Street are furry and friendly." Aside from a certain air of self-congratulation, it should be noted that "strange" is actually closer to the etymological meaning of "monster" than "evil" -- although the overtone of evil has undoubtedly crept into a significant part of modern usage. Could it be that the American Heritage editors, knowing that they would be competing with Sesame Street, decided to use their dictionary to strike a blow against monsters?

Philosophical differences are even more significant when it comes to choice of words. Both dictionaries include "robot," for example, but only M.F.D. sees fit to include a picture and definition of a robin: "A robin is a kind of bird. The front of a robin is red. Robins sing when spring begins." But then, further down the page, the S.S.D. recovers lost points by including "rock" not only as a noun (which M.F.D. also has) but as a verb: "When you rock, you move back and forth or from side to side." Will children who use M.F.D. rather than S.S.D. think the mysterious nursery rhyme means that when the wind blows the cradle will turn to stone? Will children who use S.S.D. go through life not knowing what a robin is? Fortunately for paedolexicographers as for all of us, children are more resourceful than that.

On the basis of pure inclusiveness, M.F.D. seems considerably better than S.S.D. In the sequence from "grain" to "grow," for example, it has quite a few more words: gram, grasshopper, gray, green, grew, and grin, while S.S.D. offers only two entries that are omitted by its competition: "grocery store" and "grouch." But inclusiveness is not all. The pattern seems to indicate that the people at American Heritage know more about lexicography, but the people at Sesame Street know more about kids. The probability is that kids will be more likely to want to look up "grouch" than the past tense of a verb or the name of a color. On the other hand, a dictionary that omits "gram" at this point in history is not really serious, and one that omits "grasshopper" is not really complete. The S.S.D. is to be read primarily for fun, it seems, and the inclusion of a Sesame Street cartoon, in color, to illustrate every single word reinforces this impression. M.F.D. , on the other hand, is a serious peice of work for serious-minded children. Ideally, I suppose, every child should have one of each. If I had to choose one to give to a child, I would probably choose Sesame Street and let the child have access to a larger dictionary for serious word-searching.

The best dictionary of all in this survey is the most specialized, The Oxford Junior Companion to Music. This volume should take care of the basic music-reference needs of most nonprodigies through high school. It lucidly explains the meaning of tonality, sonata form, serialism and similar arcana, discusses musical instruments ranging from the ocarina to the organ, and is particularly good on brief musical biographies. One may wonder whether the Beatles deserve nearly three times as much space as Rachmaninov (whose name is spelled "Rakhmaninov"), but the emphasis is probably right for this age-group, and, blessedly, Engelbert Humperdinck is identified as the composer of Hansel and Gretel, not as a rock singer.

The biographies include quite a few composers who would have received little or no notice in adult reference works a generation or two ago, such as Jacob Obrecht, the brothers Stamitz and the three best-known sons of J. S. Bach. American composers are well-covered from William Billings to Elliott Carter, but when it comes to the younger generation the book betrays its origins rather than Columbia or Harvard. American students, one assumes, would be likely to look up George Crumb or Charles Wuorinen (who are not included) more often than Peter Maxwell Davies or Harrison Birtwistle (who are, and should be). But this is a small blemish in a fine piece of work, as are occasional lapses in proofreading. The entry on the crumhorn, for example, is slightly misplaced (after rather than before "crush-note") but one is so happy to see it all that one hates to complain.