OPINION

On May 10, 1972, plaintiff-appellant filed a complaint in two counts against Le Conté Cosmetics, Inc., Elton C. Toland, its principal owner and executive officer, and Lenore Toland, his former wife, for infringement of plaintiff’s trademark ( 15 U.S.C. § 1051 et seq.) and unfair competition. Defendants filed an answer and counterclaims on July 10, 1972, and plaintiff replied to the counterclaims on July 27, 1972. On September 1, 1972, plaintiff asked for leave to amend its complaint to add a third claim for relief for false representations ( 15 U.S.C. § 1125 (a)). This motion was de nied. Defendants and plaintiff then both filed motions for summary judgment with accompanying affidavits. On March 26, 1973, the date set for the oral hearing on these motions, Lenore Toland was dismissed from the action by plaintiff. Defendants dismissed all of their counterclaims, and the court and parties then agreed to try the case solely on the basis of the affidavits and exhibits already presented, and certain stipulated facts, on the issue whether there was a “reasonable likelihood of confusing the source and origin of defendants’ product with that of plaintiff’s”. The court found that there was no such likelihood and entered judgment in favor of defendants. For the reasons stated herein, we reverse.

The J. B. Williams Co. has produced hand soaps and shampoos under the trademark “Conti” since 1924. Since 1966 Le Conté Cosmetics has marketed its line of cosmetics and hair care products under the mark “Le Conté”. The primary issue on this appeal is whether the trial court correctly determined there was not a likelihood of confusion of “Conti” with “Le Conté”. This Court stated in Fleischmann Distilling Corp. v. Maier Brewing Co., 314 F.2d 149, 152 (9th Cir. 1963):

See also Friend v. H. A. Friend and Co., 416 F.2d 526, 531 (9th Cir. 1969), wherein the court stated: “[L]ikelihood of confusion is a matter to be determined by this court [citing Fleischmann .

Whether likelihood of confusion is more a question of law or one of fact depends on the circumstances of each particular case. To the extent that the conclusion of the trial court is based solely upon disputed findings of fact, the appellate court must follow the conclusion of the trial court unless it finds the underlying facts to be clearly erroneous. Thus, this Court has refused on many occasions to decide de novo the fact's underlying the trial court’s determination of whether likelihood of confusion existed. See Carter-Wallace, Inc. v. Proctor & Gamble Co., 434 F.2d 794 , 799 (9th Cir. 1970); Paul Sachs Originals Co. v. Sachs, 325 F.2d 212, 214 (9th Cir. 1963); Plough, Inc. v. Kreis Laboratories, 314 F.2d 635, 641 (9th Cir. 1963). However, if the facts are not in dispute, the appellate court is “in as good a position as the trial judge to determine the probability of confusion”. Fleischmann Distilling Corp. v. Maier Brewing Co., supra, 314 F.2d at 152, quoting Miles Shoes, Inc. v. R. H. Macy & Co., 199 F.2d 602 (2d Cir. 1952).

In the present case, as in Fleischmann, no facts presented to the trial court are in dispute. The matter was submitted to the court based on affidavits, exhibits, and certain stipulat ed facts. No testimony was taken at trial. Since no issue of material fact arises from the affidavits, exhibits and stipulated facts, the determination by the trial judge, as to whether in light of those undisputed facts there existed a “likelihood of confusion” between Conti and Le Conté is a question of law, readily reviewable by this Court.

This situation is distinguished from those cases, supra, in which the trial court’s conclusion of whether or not there is a likelihood of confusion is based upon disputed issues of fact, resolved at trial. Each case of trademark infringement must be analyzed based on its own facts. The results reached in different cases decided by this circuit only appear to be contradictory when the particular facts of each case are not analyzed.

Thus, in Paul Sachs Originals Co. v. Sachs, supra, it was disputed at trial whether the name “Sachs” had become the dominant portion of appellant’s trademark. The trial court found that it had not, and a panel of this court, including two of the judges who had decided Fleischmann, supra, the same year, refused to disturb the trial court’s finding because it was not clearly erroneous.

The standard followed in this Circuit in reviewing the trial court’s decision regarding likelihood of confusion was set out in HMH Publishing Co., Inc. v. Lambert, 482 F.2d 595, 599, n. 6 (9th Cir. 1973):

To this standard, we add a corollary test, namely, where the conclusion of the trial court is based solely upon disputed findings of fact, the appellate court need not follow the conclusion of the trial court where it finds the underlying facts to be clearly erroneous.

When, as in the present case, the trial court has based its decision upon affidavits, exhibits and certain stipulated facts, none of which raises an issue of material fact, this court can, and should, determine the issue of likelihood of confusion for itself.

In order to determine whether there is a likelihood of confusion in a trademark infringement case, the Court must consider numerous factors, including inter alia the strength or weakness of the marks, similarity in appearance, sound, and meaning, the class of goods in question, the marketing channels, evidence of actual confusion, and evidence of the intention of defendant in selecting and using the alleged infringing name. See Carter-Wallace, Inc. v. Proctor & Gamble Co., supra, 434 F.2d at 800; Paul Sachs Originals Co. v. Sachs, supra, 325 F.2d at 214 . After considering these factors, the Court then must determine whether there exists a likelihood of confusion.

Structuring the analysis in this fashion, it becomes easier to determine whether the trial court findings on the listed factors are ones of law or fact. Characterization is facilitated by an example. Consider the finding in the instant case on similarity of appearance. There was no dispute as to the actual appearance of the marks. With the actual appearance on evidentiary fact, the next step was to decide whether to a reasonable viewer the marks were similar in appearance. That finding, if affirmative, would be aggregated with the other factors and the aggregate assessed as a foundation for the ultimate “likelihood of confusion” determination. Individually, each of these findings is preliminary to and not per se determinative of the ultimate issue.

Viewing the foundational question as one of “confusing similarity” is improper because it merges analysis of one of the preliminary inquiries with the conceptually distinct step of applying the statutory standard. The marks may be similar in appearance (foundational fact) yet not likely to cause confusion as to their source, particularly when all the factors are considered. Using this approach, similarity of appearance and the remaining factors provide foundational facts and should be assessed on review under the clearly erroneous rule.

The first step in the analysis is to determine whether the mark seeking protection is “strong” or “weak”. A “strong” mark is one which is used only in a “fictitious, arbitrary and fanciful manner”, see National Lead Co. v. Wolfe, 223 F.2d 195, 199 (9th Cir. 1955), whereas a “weak” mark is a mark that is a meaningful word in common usage, see Sunbeam Lighting Co. v. Sunbeam Corp., 183 F.2d 969, 972-973 (9th Cir. 1950), or is merely a suggestive or descriptive trademark, see Majestic Mfg. Co. v. Majestic Electric Appliance Co., Inc., 172 F.2d 862 (6th Cir. 1949). A “strong” mark is entitled to a greater degree of protection than is a “weak” one because of its unique usage, see Stork Restaurant v. Sahati, 166 F.2d 348, 355 (9th Cir. 1948). “Conti” must be considered a “strong” mark because it was stipulated at trial that the origin of the name is not known, and there was no evidence produced that it is a word with a meaning of its own.'

Second, the marks must be compared for similarity in appearance, sound, and meaning. See National Lead Co. v. Wolfe, supra, 223 F.2d at 201 . Here, neither mark has any clear meaning in English, but their appearances are similar. We reject the contrary finding of the trial court as clearly erroneous. Further, the trial court found, on the basis of judicial notice, that Americans would normally pronounce “Le Conté” in the French manner with the accent on the final syllable because of the heavy advertising of such brands as “Fabergé” and “Jean Naté”. That finding is not a proper subject for judicial notice.

Finally, the Court must consider the similarity in use of the products of both companies and the channels through which they are marketed. Appellees argue that the products of the two companies are so different that, even if the names are similar, there is no possibility of confusion. While the major components of some of the products of the two companies do differ, both companies produce hair care products whose uses are “related so that they are likely to be connected in the mind of a prospective purchaser”, Fleischmann Distilling Corp. v. Maier Brewing Co., supra, 314 F.2d at 159, because the products are similar and the markets for the two lines of products overlap. The trial court said in its Finding of Fact 11 that some of Conti’s products are sold in predominantly Black-American neighborhoods where, admittedly, most of Le Conté’s products are sold. This fact joins findings of similarity in the products and in the sound and appearance of the two marks plus a designation of the Conti mark as “strong”. It may be that a worldly and sophisticated person would be able to discern the differences between the marks. We conclude, however, that a reasonable customer of average intelligence and experience would very likely be confused as to the source, due to the nexus of features shared by the two marks.

We hold therefore that the court below was in error in entering judgment for appellees, because a likelihood of confusion has been shown.

A second issue raised on this appeal is whether the trial judge correctly denied appellant leave to amend its original complaint to include a claim for false representations made by appellees. The original complaint was filed on May 10, 1972. Less than six weeks after appellant claims it became aware of the basis for a third cause of action, it requested leave to amend the complaint.

While the trial judge may exercise his discretion in granting leave to amend pleadings, Foman v. Davis, 371 U.S. 178, 182 , 83 S.Ct. 227, 230 , 9 L.Ed.2d 222 (1962), Rule 15(a) of the Federal Rules of Civil Procedure provides that “leave [to amend] shall be freely given when justice so requires”, and “outright refusal to grant the leave without any justifying reason appearing for the denial is not an exercise of discretion; it is merely abuse of that discretion and inconsistent with the spirit of the Federal Rules.” 371 U.S. at 182 , 83 S.Ct. at 230 . On remand, therefore, we suggest that the trial court consider granting leave to amend the complaint.

The judgment below is reversed and the cause is remanded to the district court with directions to enter judgment in accordance with this opinion and to determine the relief to which appellant is entitled.

. See 3 R. Callman, The Law of Unfair Competition Trademarks and Monopolies, § 82.3(b) (3d ed. 1969). . In Fleischmann Distilling Corp. v. Maier Brewing Co., 196 F.Supp. 401 (N.D.Cal.1961), the trial court stated: “There is no basic dispute as to the facts, most of which have been admitted in the pleadings, by stipulations, by discovery or by uncontradicted testimony.” . It might be argued that the conflicting contentions of the parties with respect to similarity in sound and appearance of the two marks gives rise to a factual dispute. This argument is no obstacle to the scope of our review since the trial court’s determination on those questions was clearly erroneous. . Although the trial court found no evidence of actual confusion, this fact does not preclude this Court from concluding that there is a “likelihood of confusion”, as actual confusion is merely one factor to be considered by the Court when it makes its determination. . With respect to defendant’s intent the trial court found that defendants did not know of plaintiff’s name “Conti” prior to this lawsuit and that the mark “Le Conté” was derived by combining letters found in the names of defendant Toland and his wife. The Court need not decide whether these findings of the trial court are clearly erroneous, for even if the Court assumes that they are correct, the other factors discussed below require us to conclude that there is a “likelihood of confusion”. . “Confusing similarity” is another way of stating the “likelihood of confusion”. . However, this Court does not approve the “probable confusion” standard mentioned in Sunbeam, supra, 183 F.2d at 974 , insofar as it may differ from the “likelihood of confusion” standard.