DAVID A. NELSON, Circuit Judge.

This is a trademark infringement action brought under the Lanham Act, 15 U.S.C. § 1051 et seq. The plaintiff, Little Caesar Enterprises, Inc., owns two registered trademarks — the words "LITTLE CAESARS" and a drawing of a toga-clad man eating a slice of pizza — used in connection with the operation of what is said to be "America's largest carry-out pizza chain." The defendants are the past and present owners and operators of two Italian restaurants in Chattanooga, Tennessee. One of the defendants' restaurants is named "Pizza Caesar USA." Its advertising sometimes features a drawing of a horse-drawn chariot in which there is a charioteer holding a pizza pie. The district court, finding no likelihood of confusion between the plaintiff's marks and those used by the defendants, concluded that there had been no infringement. We shall affirm the judgment in favor of the defendants.

I

Plaintiff Little Caesar Enterprises operates and grants franchisees the right to operate "stores" — 90% of which are carryout stores — that sell pizza and a few other foods. The business, started in the late 1950s in Detroit, Michigan, now numbers over 1,000 outlets in 38 states, Canada and England. As of May of 1986 there were eleven Little Caesar outlets in Tennessee, and a franchisee had recently agreed to open the first such outlet in the Chattanooga area.

On January 16, 1968, plaintiff Little Caesar Enterprises had the mark "LITTLE CAESARS" registered with the United States Patent and Trademark Office, along with a "fanciful Roman" mark; both are reproduced below:

The marks attained incontestable status in 1973.

In July of 1976 defendants Caesar Randazzo and Pizza Caesar, Inc., opened an Italian restaurant in Chattanooga under the name "Pizza Caesar USA." The restaurant is operated as a conventional sitdown establishment and has very little carry-out business. In promoting his restaurant business Mr. Randazzo sometimes used the drawing and legend reproduced below:

In 1983 Mr. Randazzo sold the assets of his business (but not the shares in his corporation, Pizza Caesar, Inc.) to G.D. Enterprises, Inc., a corporation formed by Messrs. Gene Doll and William Inkster. The purchaser acquired the right to use the name "Pizza Caesar USA" for seven years.

In 1983 attorneys for plaintiff Little Caesar Enterprises decided to take action against Pizza Caesar, Inc., and Mr. Randazzo, sending them at least two certified letters objecting to the alleged trademark infringement. The recipients never responded. On April 16, 1985, Little Caesar Enterprises filed suit against Pizza Caesar, Inc., seeking injunctive relief and damages for trademark infringement. Mr. Randazzo and G.D. Enterprises were added as defendants in August of 1985.

District Judge R. Allan Edgar conducted a one-day trial on June 13, 1986. He subsequently entered judgment in favor of the defendants, finding there was no likelihood of confusion in the use of the various marks by the defendants.

II

The touchstone of trademark law is the "likelihood of confusion." Professor J. Thomas McCarthy, in Trademarks and Unfair Competition § 23:22 (2d ed. 1984), states that the law has traditionally classified likelihood of confusion as a question of fact. Some courts, however, have treated the question as one of law, particularly where "the trial court's finding rests upon an inference drawn from documentary evidence or undisputed facts." Id.

Our circuit has adopted a mixed standard of review. In Frisch's Restaurants, Inc. v. Elby's Big Boy, 670 F.2d 642 (6th Cir.), cert. denied, 459 U.S. 916, 103 S.Ct. 231, 74 L.Ed.2d 182 (1982), we listed eight factors that are to be "balanced" in determining whether there is a likelihood of confusion. In Carson v. Here's Johnny Portable Toilets, Inc., 698 F.2d 831, 833 (6th Cir. 1983), we said that "these eight foundational factors are factual and subject to a clearly erroneous standard of review, while the weighing of these findings on the ultimate issue of the likelihood of confusion is a question of law."

The eight factors listed in the Frisch's case are these:

(1) strength of the plaintiff's mark;

(2) relatedness of the goods;

(3) similarity of the marks;

(4) evidence of actual confusion;

(5) marketing channels used;

(6) likely degree of purchaser care;

(7) defendant's intent in selecting the mark;

(8) likelihood of expansion of the product lines.

See Frisch's, 670 F.2d at 648.

As to the first factor, the strength of the Little Caesars mark, it seems to us that the mark is a moderately strong one. Trademarks are generally categorized as fanciful, arbitrary, suggestive or descriptive. A "fanciful" mark is a combination of letters or other symbols signifying nothing other than the product or service to which the mark has been assigned ( e.g., Exxon, Kodak). An "arbitrary" mark has a significance recognized in everyday life, but the thing it normally signifies is unrelated to the product or service to which the mark is attached ( e.g., Camel cigarettes or Apple computers). Fanciful and arbitrary marks are considered to be the "strongest" or most distinctive marks. Encroachment on a strong mark tends to produce the greatest likelihood of confusion. "Suggestive" and "descriptive" marks either evoke some quality of the product ( e.g., Easy Off, Skinvisible) or describe it directly ( e.g., Super Glue). Such marks are considered "weaker," and confusion is said to be less likely where weak marks are involved.

The Little Caesars mark may be somewhat suggestive, to the extent that it conveys the idea that something of Italian origin is involved, but it comes closer to the "arbitrary" category. The names of historical personages are generally considered to be "arbitrary." See McCarthy, supra, at § 13.10. Insofar as the addition of the adjective "Little" to the name "Caesar" may prompt prospective pizza buyers of a certain age to think of the actor Edward G. Robinson and the motion picture in which he starred — a film that was once a well known artifact of American culture — the arbitrary character of the mark is hardly lessened.

As to the second factor, the goods and services sold by the parties in this case are quite closely related. Although it is true that Little Caesars is primarily a carryout operation and Pizza Caesar USA is primarily a sit-down restaurant, they both sell Italian food, and pizza figures more or less prominently in the menus of both. The Patent and Trademark Office has issued a "schedule of classes of goods and services" ( 37 C.F.R. § 6.1) that has no category for "restaurants" eo nomine, but the categories that are named are so broad that if there were a "restaurants" category, it seems clear that two operations as similar as those conducted by the plaintiff and the defendants would both be in it.

The third factor, the similarity of the marks, is pivotal here. Although both marks employ the word "Caesar," we are warned by Professor McCarthy that a trademark

"should not be split up into its component parts and each part then compared with parts of the conflicting mark to determine the likelihood of confusion. It is the impression which the mark as a whole creates on the average reasonably prudent buyer and not the parts thereof which is important."

McCarthy, supra, at § 23:15 (footnotes omitted). McCarthy observes that it is a violation of this "anti-dissection" rule to focus upon the "prominent" feature of a mark and not upon the mark in its totality. Thus in Long John Distilleries, Ltd. v. Sazerac Co., 426 F.2d 1406, 1407 (C.C.P.A. 1970), where the similarity between "Friar John" (brandy) and "Long John" (scotch whiskey) was held not to be confusing,

"[t]he board found that the third party registrations of record are themselves sufficient to establish that the word `John' has been so frequently used as a part of distilled beverage marks that it is incapable of indicating origin in any one source and that the substantial good will in the mark `LONG JOHN' acquired by appellant over years of use and promotion resides in the unitary mark. Further, the board was of opinion that, considered in their entireties, the differences in sound, appearance and commercial impression of the two marks are obvious and one is in no way suggestive of the other."

The word "Caesar," similarly, is frequently used with things Italian, including Italian food — viz., a type of salad and several brands of salad dressings. The differences in sound and appearance between "Little Caesar" and "Pizza Caesar" are obvious, and the addition of the acronym "USA" to the latter mark almost doubles the number of syllables and heightens the distinction. We are not unmindful, moreover, that "Caesar" is defendant Randazzo's first name, and he came to the Chattanooga market, if he did not conquer it, first. First in time is not always first in right, where trademark law is concerned, but it seems to us that priority of use may have some bearing on the likelihood of confusion.

As to the fourth Frisch's factor (actual confusion), we have no survey evidence; Little Caesar Enterprises only recently broke ground in Chattanooga, and there has been no real opportunity for actual confusion.

As far as marketing channels are concerned, both businesses are retailers, but their marketing techniques are quite different. Little Caesar Enterprises has a sophisticated advertising program that includes the use of television and that benefits, presumably, from the large number of Little Caesar outlets. Pizza Caesar USA is a small local business with few outlets and a far less sophisticated advertising program.

With regard to the sixth factor, likely degree of purchaser care, the district court thought it unlikely that the public would conduct much preprandial product research. We agree, but note that it probably makes little difference on the facts before us.

The seventh factor is the intent of the defendant. Proof of intent to appropriate another's property may be comparable to an expression of opinion by an expert witness; a defendant who purposely chooses a particular mark because it is similar to that of a senior user is saying, in effect, that he thinks there is at least a possibility that he can divert some business from the senior user — and the defendant ought to know at least as much about the likelihood of confusion as the trier of fact. See Amstar Corp. v. Domino's Pizza, Inc., 615 F.2d 252, 263 (5th Cir.), cert. denied, 449 U.S. 899, 101 S.Ct. 268, 66 L.Ed.2d 129 (1980). In the case at bar there is no evidence that Mr. Randazzo even had actual knowledge of the existence of Little Caesar Enterprises, much less that he had any larcenous intent.

The eighth factor is likelihood of expansion of the product lines. There is no need for discussion of this factor here, because the plaintiff's expansion into the product area in which the defendants have been operating is a given.

On balance, we agree with the trial court that plaintiff Little Caesar Enterprises has failed to show a likelihood of confusion between its mark and the defendant's. It is true that several of the Frisch's factors favor the plaintiff: Little Caesar has a relatively strong mark, the products and services involved are similar, and the level of consumer care is low. On the particular facts before us, however, the third factor seems more significant; we think the marks are sufficiently distinct to make confusion unlikely.

The exercise in which we have been engaging is an impressionistic one, of course, and time may prove us wrong. If actual confusion becomes common as the two businesses operate in the same area, the plaintiff will not be barred from seeking appropriate injunctive relief on the strength of evidence not in existence when this case was tried.

The judgment of the district court is AFFIRMED.