In this trademark infringement action, the district court, after a brief non-jury trial, found appellant AMF’s trade mark was valid, but not infringed, and denied AMF’s request for injunctive relief.

AMF and appellee Nescher both manufacture recreational boats. AMF uses the mark Slickcraft, and Nescher uses Sleek-craft. The crux of this appeal is whether concurrent use of the two marks is likely to confuse the public. The district judge held that confusion was unlikely. We disagree and remand for entry of a limited injunction.

1. FACTS

AMF’s predecessor used the name Slick-craft Boat Company from 1954 to 1969 when it became a division of AMF. The mark Slickcraft was federally registered on April 1, 1969, and has been continuously used since then as a trademark for this line of recreational boats.

Slickcraft boats are distributed and advertised nationally. AMF has authorized over one hundred retail outlets to sell the Slickcraft line. For the years 1966-1974, promotional expenditures for the Slickcraft line averaged approximately $200,000 annually. Gross sales for the same period approached $50,000,000.

After several years in the boat-building business, appellee Nescher organized a sole proprietorship, Nescher Boats, in 1962. This venture failed in 1967. In late 1968 Nescher began anew and adopted the name Sleekcraft. Since then Sleekcraft has been the Nescher trademark. The name Sleekcraft was selected without knowledge of appellant’s use. After AMF notified him of the alleged trademark infringement, Nescher adopted a distinctive logo and added the identifying phrase “Boats by Nescher” on plaques affixed to the boat and in much of its advertising. (See Appendix A). The Sleekcraft mark still appears alone on some of appellee’s stationery, signs, trucks, and advertisements.

The Sleekcraft venture succeeded. Expenditures for promotion increased from $6,800 in 1970 to $126,000 in 1974. Gross sales rose from $331,000 in 1970 to over .$6,000,000 in 1975. Like AMF, Nescher sells his boats through authorized local dealers.

Slickcraft boats are advertised primarily in magazines of general circulation. Nescher advertises primarily in publications for boat racing enthusiasts. Both parties exhibit their product line at boat shows, sometimes the same show.

II. ISSUES

1. What is the standard of review?

2. Has a likelihood of confusion been demonstrated?

3. What relief is warranted?

III. STANDARD OF REVIEW

Nescher contends that “likelihood of confusion” is a question of fact reviewable under the clearly erroneous standard. AMF says it is a question of law reviewable de novo. Both sides have cited case law that supports their position. The sundry formulations utilized by other courts in reviewing the same question are irreconcilable. Cf. Plough, Inc. v. Kreis Laboratories, 314 F.2d 635, 640-41 (CA 9 1963) (expresses uncertainty about what is standard of review). One commentator has severely criticized this court:

J. T. McCarthy, Trademarks and Unfair Competition § 23:23 (1973). See also United States Jaycees v. San Francisco Junior Chamber of Commerce, 354 F.Supp. 61, 77-78 (N.D.Cal.1972), aff’d, 513 F.2d 1226 (9 Cir. 1975).

Recently, however, the proper standard of review has been delineated with more clarity, and the cases in this circuit have been reconciled. In J. B. Williams Co., Inc. v. Le Conte Cosmetics, Inc., 523 F.2d 187 (CA 9 1975), cert. denied, 424 U.S. 913 , 96 5. Ct. 1110, 47 L.Ed.2d 317 (1976), we established dichotomous tests depending on whether the facts were disputed at trial:

523 F.2d at 190 (footnote in original omitted). The court then added a “corollary test”:

Id. at 191 .

Here, the facts were disputed at trial, and still are on appeal. Thus, we must first review the district court’s findings of fact, then assess its conclusion regarding the likelihood of confusion.

Before reviewing the factors relating to likelihood of confusion, we are confronted with one other argument concerning the proper standard of review. AMF argues that we should pay less deference to the district court’s findings than we ordinarily do because the court simply endorsed the findings prepared by appellee’s counsel. Although the legal proposition pressed by AMF has been recognized by some courts, it is inapposite to the procedure followed below. Before the proposed findings prepared by counsel were submitted, the district judge filed a Memorandum Opinion (R. 335) covering virtually all the issues litigated. The proposed findings are taken, often verbatim, from the court’s Memorandum Opinion. Thus, there is no reason to doubt that these findings are a product of the district court’s own deliberations to which we should defer.

IV. FACTORS RELEVANT TO LIKELIHOOD OF CONFUSION

When the goods produced by the alleged infringer compete for sales with those of the trademark owner, infringement usually will be found if the marks are sufficiently similar that confusion can be expected. When the goods are related, but not competitive, several other factors are added to the calculus. If the goods are totally unrelated, there can be no infringement because confusion is unlikely.

AMF contends these boat lines are competitive. Both lines are comprised of sporty, fiberglass boats often used for water skiing; the sizes of the boats are similar as are the prices. Nescher contends his boats are not competitive with Slick-craft boats because his are true high performance boats intended for racing enthusiasts.

The district court found that although there was some overlap in potential customers for the two product lines, the boats “appeal to separate sub-markets.” Slick-craft boats are for general family recreation, and Sleekcraft boats are for persons who want high speed recreation; thus, the district court concluded, competition between the lines is negligible. Our research has led us to only one case in which a similarly fine distinction in markets has been recognized, Sleeper Lounge Co. v. Bell Manufacturing Co., 253 F.2d 720 (CA 9 1958). Yet, after careful review of all the exhibits introduced at trial, we are convinced the district court’s finding was warranted by the evidence.

The Slickcraft line is designed for a variety of activities: fishing, water skiing, pleasure cruises, and sunbathing. The promotional literature emphasizes family fun. Sleekcraft boats are not for families. They are low-profile racing boats designed for racing, high speed cruises, and water skiing. Seating capacity and luxury are secondary. Unlike the Slickcraft line, handling capability is emphasized. The promotional literature projects an alluring, perhaps flashy, racing image; absent from the pictures are the small children prominently displayed in the Slickcraft brochures.

Even though both boats are designed for towing water skiers, only the highly skilled enthusiast would require the higher speeds the Sleekcraft promises. We therefore affirm the district court’s finding that, despite the potential market overlap, the two lines are not competitive. Accordingly, we must consider all the relevant circumstances in assessing the likelihood of confusion. See Durox Co. v. Duron Paint Manufacturing Co., 320 F.2d 882, 885 (CA 4 1963).

V. FACTORS RELEVANT TO LIKELIHOOD OF CONFUSION

In determining whether confusion between related goods is likely, the following factors are relevant:

See, e. g., Sleeper Lounge Co. v. Bell Manufacturing Co., 253 F.2d at 722 ; Restatement of Torts § 731 (1938). We discuss each serially.

1. Strength of the mark

A strong mark is inherently distinctive, for example, an arbitrary or fanciful mark; it will be afforded the widest ambit of protection from infringing uses. See, e. g., National Lead Co. v. Wolfe, 223 F.2d 195, 199 (CA 9), cert. denied, 350 U.S. 883 , 76 S.Ct. 135 , 100 L.Ed. 778 (1955) (Dutch Boy not used geographically or descriptively, but in a “fictitious, arbitrary and fanciful manner”). A descriptive mark tells something about the product; it will be protected only when secondary meaning is shown. See Miss Universe, Inc. v. Patricelli, 408 F.2d 506 (CA 2 1969); cf. Hesmer Foods, Inc. v. Campbell Soup Co., 346 F.2d 356 (CA 7), cert. denied, 382 U.S. 839 , 86 S.Ct. 89 , 15 L.Ed.2d 81 (1965) (barbecue beans used as a description, not a trademark). In between lie suggestive marks which subtly connote something about the products. Although less distinctive than an arbitrary or fanciful mark and therefore a comparatively weak mark, a suggestive mark will be protected without proof of secondary meaning. Watkins Products, Inc. v. Sunway Fruit Products, Inc., 311 F.2d 496 (CA 7 1962).

Slickcraft is, AMF asserts, a fanciful mark and therefore entitled to wide protection. This assertion is incorrect. The issue, as we view it, is whether Slickcraft is descriptive or suggestive of appellant’s boats.

The district court did not make any explicit finding regarding the strength of appellant’s mark. Implicitly, however, the findings indicate the mark was viewed as suggestive: proof of secondary meaning was not offered or discussed, and yet the district court concluded that the mark Slickcraft was valid and deserved some protection from potential infringement.

Whether Slickcraft is suggestive or descriptive is a close question. The line separating the two is uncertain; extrapolating the line from precedent would be impossible. Compare Van Camp Sea Food Co. v. Alexander B. Stewart Organizations, 50 F.2d 976, 979 (Cust. & Pat.App.1931) with Van Camp Sea Food Co. v. Packman Bros., 4 F.Supp. 522 (D.N.J.1933), aff’d 79 F.2d 511 (3 Cir. 1935).

Although the distinction between descriptive and suggestive marks may be inarticulable, several criteria offer guidance. See generally McCarthy, supra, at §§ 11:21-11:22. The primary criterion is “the imaginativeness involved in the suggestion,” Restatement of Torts § 721, Comment a (1938): that is, how immediate and direct is the thought process from the mark to the particular product. From the word Slickcraft one might readily conjure up the image of appellant’s boats, yet a number of other images might also follow. A secondary criterion is whether granting the trademark owner a limited monopoly will in fact inhibit legitimate use of the mark by other sellers. There is no evidence here that others have used or desire to use Slickcraft in describing their goods. Another criterion is whether the mark is actually viewed by the public as an indication of the product’s origin or as a self-serving description of it. See Carter-Wallace, Inc. v. Procter & Gamble Co., 434 F.2d 794 (CA 9 1970). We think buyers probably will understand that Slick-craft is a trademark, particularly since it is generally used in conjunction with the mark AMF. (See Appendix A). Based on the above criteria and our reading of the district court’s findings, we hold that Slick-craft is a suggestive mark when applied to boats. But see Norm Thompson Outfitters, Inc. v. General Motors Corp., 448 F.2d 1293, 1295 (CA 9 1971) (applying Oregon law, court upholds finding by district court that slogan “Escape From The Ordinary” descriptive of assortment of unusual items).

Although appellant’s mark is protective and may have been strengthened by advertising, see Carter-Wallace, Inc. v. Procter & Gamble Co., 434 F.2d at 800 ; Stork Restaurant, Inc. v. Sahati, 166 F.2d 348, 356 (CA 9 1948), it is a weak mark entitled to a restricted range of protection. Thus, only if the marks are quite similar, and the goods closely related, will infringement be found. See Westward Coach Manufacturing Co. v. Ford Motor Co., 388 F.2d 627, 634 (CA 7), cert. denied, 392 U.S. 927 , 88 S.Ct. 2286 , 20 L.Ed.2d 1386 (1968) (weak mark protected only against similar goods, similarly marketed).

2. Proximity of the goods

For related goods, the danger presented is that the public will mistakenly assume there is an association between the producers of the related goods, though no such association exists. Cf. Stork Restaurant, Inc. v. Sahati, 166 F.2d at 356 (restaurants with same name in different geographical and price markets). The more likely the public is to make such an association, the less similarity in the marks is requisite to a finding of likelihood of confusion. See American Steel Foundries v. Robertson, 269 U.S. 372, 382 , 46 S.Ct. 160 , 70 L.Ed. 317 (1926); Russell Chemical Co. v. Wyandotte Chemicals Corp., 337 F.2d 660 , 52 C.C.P.A. 807 (1964). Thus, less similarity between the marks will suffice when the goods are complementary, Communications Satellite Corp. v. Comcet, Inc., 429 F.2d at 1253 (“Complementary products, or services, are particularly vulnerable to confusion.”); American Drill Bushing Co. v. Rockwell Manufacturing Co., 342 F.2d 1019, 1022 , 52 C.C.P.A. 1173 (1965) (drill bits and drill bushings), the products are sold to the same class of purchasers, Restatement of Torts § 731, Comment d (1938), or the goods are similar in use and function. See Syntex Laboratories, Inc. v. Norwich Pharmacal Co., 437 F.2d 566, 569 (CA 2 1971).

Although these product lines are non-competing, they are extremely close in use and function. In fact, their uses overlap. Both are for recreational boating on bays and lakes. Both are designed for water skiing and speedy cruises. Their functional features, for the most part, are also similar: fiberglass bodies, outboard motors, and open seating for a handful of people. Although the Sleekcraft boat is for higher speed recreation and its refinements support the market distinction the district court made, they are so closely related that a diminished standard of similarity must be applied when comparing the two marks. See Standard Brands, Inc. v. Smidler, 151 F.2d at 37 (vegetable juice and vitamin tablets “closely related”); cf. Russell Chemical Co. v. Wyandotte Chemicals Corp., 337 F.2d at 661 (when marks, i. e., Sentol and Sen-trol, nearly identical in sight and sound, products must be “substantially different” for confusion to be unlikely).

3. Similarity of the marks

The district court found that “the two marks are easily distinguishable in use either when written or spoken.” Again, there is confusion among the cases as to whether review of this finding is subject to the clearly erroneous standard.

In resolving this dispute, we must be mindful of exactly what the district court determined. Although in some cases similarity of the marks has been equated with likelihood of confusion, the two inquiries are separate:

J. B. Williams Co., Inc. v. LeConte Cosmetics, Inc., 523 F.2d at 192 . Thus, as in J. B. Williams Co., even when the evidence is not in conflict, the district court’s findings on similarity of the marks must be upheld unless clearly erroneous.

Similarity of the marks is tested on three levels: sight, sound, and meaning. See Plough, Inc. v. Kreis Laboratories, 314 F.2d 635, 638 (CA 9 1963). Each must be considered as they are encountered in the marketplace. Although similarity is measured by the marks as entities, similarities weigh more heavily than differences. Esso Standard Oil Co. v. Sun Oil Co., 97 U.S.App.D.C. 154 , 157, 229 F.2d 37, 40 , cert. denied, 351 U.S. 973 , 76 S.Ct. 1027 , 100 L.Ed. 1491 (1956).

Standing alone the words Sleek-craft and Slickcraft are the same except for two inconspicuous letters in the middle of the first syllable. See Communications Satellite Corp. v. Comcet, Inc., 429 F.2d at 1249 (only slight visual difference between Com-sat and Comcet); Polaroid Corp. v. Polaraid, Inc., 319 F.2d 830 (CA 7 1963) (Polaroid cameras and lenses and Polaraid heating and refrigeration systems); cf. Drexel Enterprises, Inc. v. Hermitage Cabinet Shop, Inc., 266 F.Supp. 532, 538 (N.D.Ga.1967) (addition of single letter on name of furniture line does not change appearance). To the eye, the words are similar.

In support of the district court’s finding, Nescher points out that the distinctive logo on his boats and brochures negates the similarity of the words. We agree: the names appear dissimilar when viewed in conjunction with the logo, but the logo is often absent. The exhibits show that the word Sleekcraft is frequently found alone in trade journals, company stationery, and various advertisements.

Nescher also points out that the Slick-craft name is usually accompanied by the additional trademark AMF. As a result of this consistent use, Nescher argues, AMF has become the salient part of the mark indicative of the product’s origin. See Rockwood Chocolate Co. v. Hoffman Candy Co., 372 F.2d 552, 555 , 54 C.C.P.A. 1061 (1967) (likelihood of confusion found for Cup-o-Gold and Bag-o-Gold, but not for Rockwood Bag-o-Gold); cf. John Morrell & Co. v. Reliable Packing Co., 295 F.2d 314 (CA 7 1961) (both parties used house name to identify origin of prepared hams).

Although Nescher is correct in asserting that use of a housemark can reduce the likelihood of confusion, Ye Olde Tavern Cheese Products, Inc. v. Planters Peanuts Division, Standard Brands, Inc., 261 F.Supp. 200, 205-06 (N.D.Ill.1966), aff’d per curiam, 394 F.2d 833 (7 Cir. 1967), the effect is negligible here even though AMF is a well-known house name for recreational equipment. The exhibits show that the AMF mark is down-played in the brochures and advertisements; the letters AMF are smaller and skewed to one side. Throughout the promotional materials, the emphasis is on the Slickcraft name. Accordingly, we find that Slickcraft is the more conspicuous mark and serves to indicate the source of origin to the public. See Celanese Corp. v. DuPont De Nemours & Co., 154 F.2d 143 , 33 C.C.P.A. 857 (1946) (DuPont the more conspicuous part of mark, but consumers more likely to remember Clar-apel as the indicator of origin; thus, Clar-apel is dominant part of DuPont Clar-apel mark).

Another argument pressed by Nescher is that we should disregard the common suffix “craft” and compare Slick and Sleek alone. Although these are the salient parts of the two marks, we must consider the entire mark. Craft, a generic frequently used in trademarks on boats, is not itself protectible, yet the common endings do add to the marks’ similarity. See Brown-Forman Distillery Co. v. Arthur M. Bloch Liquor Importers, Inc., 99 F.2d 708, 710 (CA 7 1938) (“Old” itself cannot be infringed, but can be considered in assessing similarity of Old Forester and Old Foster). The difference between Slick and Sleek is insufficient to overcome the overall visual similarity.

Sound is also important because reputation is often conveyed word-of-mouth. We recognize that the two sounds can be distinguished, but the difference is only in a small part of one syllable. In G. D. Searle & Co. v. Chas. Pfizer & Co., 265 F.2d 385 (CA 7 1959), cert. denied, 361 U.S. 819 , 80 S.Ct. 64 , 4 L.Ed.2d 65 (1959), the court reversed the trial court’s finding that Bonamine sounded “unlike” Dramamine, stating that: “Slight differences in the sound of trademarks will not protect the infringer.” Id. at 387. The difference here is even slighter. In a case in which the difference in sound corresponds closely with the instant appeal, the Court of Customs and Patent Appeals held that the sounds Pediglo and Pechglo were similar. Vanity Fair Mills, Inc. v. Pedigree Fabrics, Inc., 161 F.2d 226, 228 , 34 C.C.P.A. 1043 (1947). In reversing the patent commissioner’s decision allowing the Pediglo mark to be registered, the court said that the question was not a “close one.” Id. See also Celanese Corp. v. DuPont De Nemours & Co., 154 F.2d at 145 , 33 C.C.P.A. 857 (Clar-apel and Clarifoil “sound very nearly alike”); David Sherman Corp. v. Heublein, Inc., 340 F.2d 377, 381 (CA 8 1965) (Smirnoff and Sarnoff “strikingly alike when spoken”).

Neither expert testimony nor survey evidence was introduced below to support the trial court’s finding that the marks were easily distinguishable to the eye and the ear. Compare G. D. Searle & Co., supra (expert testimony offered on phonetic similarity). The district judge based his conclusion on a comparison of the marks. After making the same comparison, we are left with a definite and firm conviction that his conclusion is incorrect. See, e. g., Syntex Laboratories, Inc. v. Norwich Pharmacal Co., 437 F.2d 566, 569 (CA 2 1971) (Vagitrol and Vagestrol).

The final criterion reinforces our conclusion. Closeness in meaning can itself substantiate a claim of similarity of trademarks. See, e. g., S. C. Johnson & Son, Inc. v. Drop Dead Co., 210 F.Supp. 816 (S.D.Cal. 1962), aff’d, 326 F.2d 87 (9 Cir. 1963) (Pledge and Promise). Nescher contends the words are sharply different in meaning. This contention is not convincing; the words are virtual synonyms. Webster’s New World Dictionary of the American Language 1371 (1966).

Despite the trial court’s findings, we hold that the marks are quite similar on all three levels. See National Lead Co. v. Wolfe, 223 ■F.2d at 201.

4. Evidence of actual confusion

Evidence that use of the two marks has already led to confusion is persuasive proof that future confusion is likely. Plough, Inc. v. Kreis Laboratories, 314 F.2d at 639 . Proving actual confusion is difficult, however, G. D. Searle & Co. v. Chas. Pfizer & Co., 265 F.2d at 389 , and the courts have often discounted such evidence because it was unclear or insubstantial. E. g., Carter-Wallace, Inc. v. Procter & Gamble Co., 434 F.2d at 800 (letters did not show actual confusion); cf. Norm Thompson Outfitters, Inc. v. General Motors Corp., 448 F.2d at 1295 (showing insufficient to establish secondary meaning).

AMF introduced evidence that confusion had occurred both in the trade and in the mind of the buying public. A substantial showing of confusion among either group might have convinced the trial court that continued use would lead to further confusion. See Continente v. Continente, 378 F.2d 279, 281-82 (CA 9 1967).

The district judge found that in light of the number of sales and the extent of the parties’ advertising, the amount of past confusion was negligible. We cannot say this finding is clearly erroneous though we might have viewed the evidence more generously.

Because of the difficulty in garnering such evidence, the failure to prove instances of actual confusion is not dispositive. Drexel Enterprises, Inc. v. Hermitage Cabinet Shop, Inc., 266 F.Supp. at 537 . Consequently, this factor is weighed heavily only when there is evidence of past confusion or, perhaps, when the particular circumstances indicate such evidence should have been available.

5. Marketing channels

Convergent marketing channels increase the likelihood of confusion. McCarthy, supra, at § 24.6. There is no evidence in the record that both lines were sold under the same roof except at boat shows; the normal marketing channels used by both AMF and Nescher are, however, parallel. Each sells through authorized retail dealers in diverse localities. The same sales methods are employed. The price ranges are almost identical. Each line is advertised extensively though different national magazines are used; the retail dealers also promote the lines, by participating in smaller boat shows and by advertising in local newspapers and classified telephone directories. Although different submarkets are involved, the general class of boat purchasers exposed to the products overlap.

6. Type of goods and purchaser care

Both parties produce high quality, expensive goods. According to the findings of fact, the boats “are purchased only after thoughtful, careful evaluation of the product and the performance the purchaser expects.”

In assessing the likelihood of confusion to the public, the standard used by the courts is the typical buyer exercising ordinary caution. See HMH Publishing Co. v. Lambert, 482 F.2d 595 , 599 n. 6 (CA 9 1973). Although the wholly indifferent may be excluded, McCarthy, supra, at § 23:27, the standard includes the ignorant and the credulous. Stork Restaurant, Inc. v. Sahati, 166 F.2d at 358 ; see Drexel Enterprises, Inc. v. Hermitage House Furniture, Inc., 309 F.Supp. 1389, 1390 (D.S.C. 1970). When the buyer has expertise in the field, a higher standard is proper though it will not preclude a finding that confusion is likely. American Drill Bushing Co. v. Rockwell Manufacturing Co., 342 F.2d at 1022 , 52 C.C.P.A. 1173 (expertise in field differentiated from expertise as to trademarks). Similarly, when the goods are expensive, the buyer can be expected to exercise greater care in his purchases; again, though, confusion may still be likely. See Omega Importing Corp. v. Petri-Kine Camera Co., 451 F.2d 1190, 1195 (CA 2 1971) (warns against undue reliance on the “supposed sophistication” and care of consumers when expensive goods are involved).

The parties vigorously dispute the validity of the trial court’s finding on how discriminating the average buyer actually is. Although AMF presented expert testimony to the contrary, the court’s finding is amply supported, by the record. The care exercised by the typical purchaser, though it might virtually eliminate mistaken purchases, does not guarantee that confusion as to association or sponsorship is unlikely.

The district court also found that trademarks are unimportant to the average boat buyer. Common sense and the evidence indicate this is not the type of purchase made only on “general impressions.” See Fleischmann Distilling Corp. v. Maier Brewing Co., 314 F.2d at 161 . This inattention to trade symbols does reduce the possibilities for confusion. Restatement of Torts § 731(g).

The high quality of defendant’s boats is also relevant in another way. The hallmark of a trademark owner’s interest in preventing use of his mark on related goods is the threat such use poses to the reputation of his own goods. See Triumph Hosiery Mills, Inc. v. Triumph International Corp., 308 F.2d at 197. When the alleged infringer’s goods are of equal quality, there is little harm to the reputation earned by the trademarked goods. Yet this is no defense, for present quality is no assurance of continued quality. See Aunt Jemima Mills Co. v. Rigney & Co., 2 Cir., 247 F. 407 at 410. The wrong inheres in involuntarily entrusting one’s business reputation to another business; see Ross-Whitney Corp. v. Smith, Kline & French Laboratories, 207 F.2d 190, 197 (CA 9 1953); AMF, of course, cannot control the quality of Sleekcraft boats. In addition, what may be deemed a beneficial feature in a racing boat may be seen as a deficiency to a person seeking a craft for general-purpose recreation; the confused consumer may then decide, without even perusing one, that a Slickcraft boat will not suit his needs. Finally, equivalence in quality may actually contribute to the assumption of a common connection.

7. Intent

The district judge found that Nescher was unaware of appellant’s use of the Slick-craft mark when he adopted the Sleekcraft name. There was no evidence that anyone attempted to palm off the latter boats for the former. And after notification of the purported infringement, Nescher designed a distinctive logo. (See Appendix A) Cf. Drop Dead Co. v. S. C. Johnson & Son, Inc., 326 F.2d at 90 (defendant modified label after action filed, but continued using infringing trademark). We agree with the district judge: appellee’s good faith cannot be questioned.

When the alleged infringer knowingly adopts a mark similar to another’s, reviewing courts presume that the defendant can accomplish his purpose: that is, that the public will be deceived. Fleischmann Distilling Corp. v. Maier Brewing Co., 314 F.2d 157 -58; see Friend v. H. A. Friend & Co., 416 F.2d 526 (CA 9 1969) cert. denied, 397 U.S. 914 , 90 S.Ct. 916 , 25 L.Ed.2d 94 (1970); cf. International Luggage Registry v. Avery Products Corp., 541 F.2d 830 (CA 9 1976) (copyright action). Good faith is less probative of the likelihood of confusion, yet may be given considerable weight in fashioning a remedy.

8. Likelihood of expansion

Inasmuch as a trademark owner is afforded greater protection against competing goods, a “strong possibility” that either party may expand his business to compete with the other will weigh in favor of finding that the present use is infringing. Restatement of Torts § 731(b) & Comment c. When goods are closely related, any expansion is likely to result in direct competition. Compare Communications Satellite Corp. v. Comcet, Inc., 429 F.2d at 1253. The evidence shows that both parties are diversifying their model lines. The potential that one or both of the parties will enter the other’s submarket with a competing model is strong.

VI. REMEDY

Based on the preceding analysis, we hold that Nescher has infringed the Slick-craft mark. Since Neseher’s use is continuing, an injunction should have been entered. Several considerations, however, convince us that a limited injunction will suffice.

Both parties have used their trademarks for over a decade. AMF has a substantial investment in the Slickcraft name, yet a complete prohibition against appellee’s use of the Sleekcraft name is unnecessary to protect against encroachment of appellant’s mark, see Miss Universe, Inc. v. Patricelli, 408 F.2d at 510 , or to eliminate public confusion. Appellee has also expended much effort and money to build and preserve the goodwill of its mark.

There is little doubt both parties honestly desire to avoid confusion of their products. Nescher adopted the Sleekcraft name in good faith and has subsequently taken steps to avoid confusion. Use of the Nescher logo in all facets of the business would ensure that confusion would not occur. The exhibits, particularly the yellow pages advertisements, convince us that this is' not being done.

Thus, in “balancing the conflicting interests both parties have in the unimpaired continuation of their trade-mark use,” Du-rox Co. v. Duron Paint Manufacturing Co., 4 Cir., 320 F.2d 882 at 885 , and the interest the public and the trade have in avoiding confusion, we conclude that a limited mandatory injunction is warranted. Upon re mand the district court should consider the above interests in structuring appropriate relief. At minimum, the logo should appear in all advertisements, signs, and promotional materials prepared either by appellee or by his retail dealers, and on all appellee’s business forms except those intended for strictly internal use. A specific disclaimer of any association with AMF or the Slick-craft line seems unnecessary, nor do we think it necessary to enjoin Nescher from expanding his product line. In its discretion the district judge may allow appellee sufficient time to consume supplies at hand and to add the logo to more permanent assets, such as business signs.

REVERSED and REMANDED.

Appendix A

. The complaint alleged two other theories of recovery, unfair competition and dilution. Under these facts, the unfair competition count adds nothing. Appellant’s dilution theory was not proved at trial: the Slickcraft mark was not shown to be strong enough to be diluted, nor did the evidence show that use of appellee’s mark on his boats would tarnish the Slickcraft image. Thus, we need not decide whether California recognizes a cause of action for dilution. See generally HMH Publishing Co. v. Lambert, 482 F.2d 595, 599 (CA 9 1973). . AMF also sought an accounting. On appeal AMF has effectively abandoned its request for an accounting. . Sleekcraft Boats, wholly owned by Nescher, is also a party-appellee. We refer to both as Nescher. . The parties and the district court assumed this gave AMF absolute priority, for appellee never contended he was entitled to protection as a good faith intermediate user — one who innocently adopts an infringing mark after the trademark owner begins to use his mark, but before he registers it. This can be a limited geographical defense. Burger King of Florida, Inc. v. Hoots, 403 F.2d 904 (CA 7 1968); 15 U.S.C. § 1115 (b); see Hanover Star Milling Co. v. Metcalf, 240 U.S. 403 , 36 S.Ct. 357 , 60 L.Ed. 713 (1916). It was appellee’s burden to raise and prove this defense, and he has not. . Our disposition obviates the need to discuss - other issues raised by appellant. “Compare: Fleischmann Distilling Corp. v. Maier Brewing Co. (1963, CA9) 314 F.2d 149 (Court of Appeals can reweigh decision on confusion) with Norm Thompson Outfitters, Inc. v. General Motors Corp. (1971, CA9 Or.) 448 F.2d 1293 (Court of Appeals must adhere to clearly erroneous rule).” . We do not understand this to mean that our hands are tied if the court’s conclusion is wholly with its incompatible findings. . We recognize that commentators have questioned this rule. J. T. McCarthy, Trademarks and Unfair Competition § 23:22 (1973); see Wright and Miller, Federal Practice and Procedure § 2587 (1971). Perhaps, it might be more accurate to say that the trial court’s errors are always clear to us when the evidence is not in dispute. In such cases adherence to the clearly erroneous formulation may be no more than semantics. Cf. Miller v. United States, 587 F.2d 991 (CA 9 1978) (“It follows in most cases where the trial judge has erred in determining what standard of conduct should have been used in a negligence determination, that the ultimate finding as to negligence does not pass muster under the ‘clearly erroneous’ test.”) . Wright & Miller, supra note 7 , at § 2578. . The alleged infringer’s intent in adopting the mark is weighed, both as probative evidence of the likelihood of confusion and as an equitable consideration. . Related goods are those “products which would be reasonably thought by the buying public to come from the same source if sold under the same mark.” Standard Brands, Inc. v. Smidler, 151 F.2d 34, 37 (CA 2 1945). See . The list is not exhaustive. Other variables may come into play depending on the particular facts presented. Triumph Hosiery Mills, Inc. v. Triumph International Corp., 308 F.2d 196, 198 (CA 2 1962); Restatement of Torts § 729, Comment a (1938). . Once secondary meaning is established, though, the protection afforded should be commensurate with the degree of consumer association proven. See Communications Satellite Corp. v. Comcet, Inc., 429 F.2d 1245, 1248 (CA 4 1970). “ ‘Confusing similarity’ is another way of stating the ‘likelihood of confusion’.” . We recognize that actions to register a recently-adopted trademark are distinguishable from infringement actions because the applicant for registration has an infinity of other marks to choose from. Consequently, in registration cases any doubts about the likelihood of confusion will be resolved against the party seeking to register a new mark. Esso Standard Oil Co. v. Sun Oil Co., 97 U.S.App.D.C. 154 , 159, 229 F.2d 37, 42 (CADC), cert. denied, 351 U.S. 973 , 76 S.Ct. 1027 , 100 L.Ed. 1491 (1956).