THE SUPREME COURT OF APPEAL OF SOUTH AFRICA
JUDGMENT
Not Reportable
Case no: 759/2023
In the matter between:
CASADOBE PROPS 60 (PTY) LTD APPELLANT
and
FRATELLI MARTINI SECONDO LUIGI S pA RESPONDENT
Neutral citation: Casadobe Props 60 (Pty) Ltd v Fratelli Martini Secondo
Luigi Sp A (Case no 759/2023) [202 5] ZASCA 14
(25 February 2025 )
Coram: DAMBUZA, MOKGOHLOA and NICHOLLS JJA and DOLAMO
and NAIDOO AJJA
Heard : 13 September 2024
Delivered : 25 February 2025
Summary: Trade Mark – opposition to the registration of a trade mark in
reliance upon s 10(14) of the Trade Marks Act 194 of 1993 – whether the trade
marks are confusingly or deceptively similar – improbable that the public would
be deceived or confused.
2
ORDER
__ ___
On appeal from: Western Cape Division of the High Court, Cape Town (Nyati
AJ sitt ing as court of first instance):
1 The appeal is upheld with costs , including the costs of two counsel.
2 The order of the high court is set aside and replaced with the following :
‘The application is dismissed with costs .’
JUDGMENT
__ ___
Mokgohloa JA ( Dambuza and Nicholls JJA and Dolamo and Naidoo AJJA
concurring):
[1] The appellant, Casadobe Props 60 (Pty) Ltd, (Casadobe), appeals a
decision of the Western Cape Division of the High Court (the high court), which
granted the respondent, F ratelli Martini Secondo Luigi Sp A (Fratelli), an interdict
restraining Casadobe from infringing Fratelli’s rights acquired by the registration
of the Trade Mark No 2012/08843 CANTI in class 33, by using the mark CANTO
or any other mark nearly resembling or so similar thereto as to be likely to cause
deception or confusion in terms of s 34(1)(a) of the Trade Marks Act 194 of 1993
(the Act) in the course of trade and in relation to wines and sparkling wines. The
appeal is with leave of this Court.
3
[2] The issue in this appeal is whether the trade mark is likely to confuse or
deceive as contemplated in s 10(14) of the Act. This section provides:
‘10 Unregistrable trade marks .-
The following marks shall not be registered as trade marks or, if registered , shall, subject to the
provisions of section 3 and 70, be liable to be removed from the register:
. . .
Subje ct to the provisions of section 14, a mark which is identical to a registered trade mark
belonging to a different proprietor or so similar thereto that the use thereof in relation to goods
or services in respect of which it is sought to be registered and which are the same or so similar
to the goods or services in respect of which such trade mark is reg istered, would be likely to
deceive or cause confusion, unless the proprietor of such trade mark consents to the registration
of such mark.’
The facts
[3] Casadobe is a South African company wh ich owns and operates the Canto
Wines boutique wine estate in the Durbanvillle wine valley, Western Cape. It
uses the unregistered trade mark CANTO to market its wine . Fratelli is an Italian
company with its principal place of business in C ossano Belbo , Italy. It is the
registered proprietor of a South African trade mark registration number
2012/08843 CANTI in class 33 wines and sparkling wines.
[4] In April 2015, Casadobe conducted a trade mark search for CANTO in
relation to wine. The search revealed ‘various ’ cases in class 33 but the Registrar
of Trade Marks did not regard the numerous trade marks incorporat ing CANT as
relevant . Fratelli ’s registration of CANTI was not cited. Based on this, and on 24
April 2020, Casadobe proceeded with its application for registration of the mark
CANTO in various classes including class 33 under registration number
2015/10717. During the examin ation of Casadobe’s application for registration
4
of its CANTO mark, the Registrar of Trademarks only cited trade mark
registration number 2007/00570 in class 33 .
[5] The other trade mark applications which have been filed in class 33 at the
trade mark ’s registry incorporating CANT included CANTANA, CANTON,
CANTI and CLUB DE CANTINEROS. These trade marks co -exist on the
registry. The re are two further trade mark applications including the letters CANT
which had been advertised by the Registrar of Trademarks.
[6] The trade mark search further revealed older trad e mark registration s in
class 33 which included the letters CANT but which had expired due to non -
payment of renewal fees. These include d CANTAMELO, CANTEVAL and
CANTORI, which co -existed with CANTA TA, CANTO N, CANTI, CLUB DE
CANTINER OS, DE CANTATA and CANTULA.
[7] In granting the interdict , the high court found that the trade marks CANTO
and C ANTI are visually, aurally and conceptually similar. It considered that both
trade marks consist of five letters with the first four letters overlapping directly,
and they have syllables CAN -TO and CAN -TI which look and sound the same.
Another similarity that the high court found was that the trade marks have a
similar meaning which is associated with a song or the act of singing .
Consequently, the high court held that the single letter difference of the single
letters ‘O’ and ‘I’ d id not sufficiently distinguish the trade marks from each other
and that there was therefore a high likelihood th at the marks may confuse
customers.
5
In this Court
[8] Section 10(14) of the Act prohibits registration of, amongst others, a mark
that is so similar to a registered mark that the use thereof in relation to the goods
or services in respect of which it is sought to be registered, which are the same or
similar to the goods or services in respect of which the opponent’s mark is
registered, ‘would be likely to deceive or cause confusion’.
[9] In considering whether the use of the appellant’s trade mark is likely to
deceive or cause confusion, this Court in Cowbell AG v ICS Holdings Ltd1 held
that the essential function of the trade mark is to indicate the origin of the goods
in connection with which it is used . The decision whether there is a reasonable
likelihood of confusion is a value judgment. In Orange Brand Services Limited v
Account Works Software (Pty) Ltd (Orange Brand )2 it was stated that in
opposition proceedings the question that falls to be decided is not how the parties
use or intend to use their marks, but how they would be entitled to use them if
both of them are registered; that is, how they might notionally be used.
[10] In Orange Brand, this Court, citing the authors of Kerly’s Law of Trade
Marks and Trade Names ,3 pointed out that the question of whether there is
likelihood of confusion or deception is a question of fact and that decided cases
in relation to other facts are of little assistance, save insofar as they lay down any
general principle. In Roodezandt Ko -operatiewe Wynmakery Ltd v Robertson
Winery (Pty) Ltd & Another4 this Court elaborated on the meaning of the value
1 Cowbell AG v ICS Holdings Ltd 2001 (3) SA 941 (SCA); [2001] 4 All SA 242 (A)
para 10 .
2 Orange Brand Services Limited v Account Works Software (Pty) Ltd [2013 ] ZASCA 158. (SCA); 2013 JDR
2750 (SCA) para 6.
3 David Kitchin et al Kerly’s Law of Trade Marks and Trade Names 14 ed (2011) para 17 – 023.
4 Roodezandt Ko -operatiewe Wynmakery Ltd v Robertson Winery (Pty) Ltd & Another [2014] ZASCA 173 ; 2014
JDR 2452 (SCA) .
6
judgment to be made , in the context of an application to remove a trade mark
from the register as follows:
‘Considerations that could assist in the exercise of this value judgment have been proposed in
numerous decided cases. One of these cases is Laboratoire Lachartre SA v Armour -Dial
Incorporated 1976 (2) SA 744 (T) in which Colman J sounded the following note of caution
(at 746B -E):
“We have had ample time for full consideration and close comparison of the two trademarks
with which we are concerned. These advantages, however, carry their own dangers. They have
caused us to look at the trademarks with far greater care than they would be looked at by the
members of the public whose probable reactions we are required to assess, and with a far keener
awareness of similarities and dissimilarities than such people would probably have as they go
about their daily lives.
What we have now to do is, therefore, to transport ourselves, notionally, from the court -room
or the study, to the market place. We must try to look at the marks as they will be seen, if they
are both in fair and normal commercial use, by the hypothetical consumers of [wine]. Those
will be people of many races and degrees of education, having varied gifts, interests and talents.
We are not to postulate the consumer of “phenomenal ignorance or extraor dinarily defective
intelligence” . . . We are to consider a person of average intelligence and proper eyesight,
buying with ordinary caution ”.’5
These principles apply with equal force to opposition proceedings. The value
judgment is likely a matter of first impression and there should not be undue -
peering at the two marks to find similarities or differences.6
[11] In Century City Apartments Property Services CC and Another v Century
City Property Owners’ Association7, this Court referred with approval to what
was stated in Compass Publishing BV v Compass Logistics Ltd8 that:
5 Ibid para 5 .
6 Yuppiechef Holdings (Pty) Ltd v Yuppie Gadgets Holdings (Pty) Ltd [2016] ZASCA 118; 2016 JDR 1705
(SCA) para 26 .
7 Century City Apartments Property Services CC and Another v Century City Property Owners’ Association
2009 BIP 230 (SCA); 2010 (3) SA 1 (SCA); [2009] ZASCA 157 para13.
8 Compass Publishing BV v Compass Logistics Ltd [2004] EWHC 520 (Ch); [2004] RPC 41 para 24 .
9 Op cit 7 para 13.
7
‘The likelihood of confusion must be appreciated globally, taking account of all relevant
factors. It must be judged through the eyes of the average consumer of the goods or services in
question. That customer is to be taken to be reasonably well informed a nd reasonably
circumspect and observant, but he may have to rely upon an imperfect picture or recollection
of the marks. The court should factor in the recognition that the average consumer normally
perceives a mark as a whole and does not analyse its vari ous details. The visual, aural and
conceptual similarities of the marks must be assessed by reference to the overall impressions
created by the marks bearing in mind their distinctive and dominant components. Furthermore,
if the association between the mar ks causes the public to wrongly believe that the respective
goods come from the same or economically linked undertakings, there is a likelihood of
confusion. ’9
[12] The test to determine whether the two marks are confusingly similar is well
established and was set out by this Court in Plascon - Evans Paints (Pty) Ltd v
Van Riebeeck Paints (Pty) Ltd (Plascon Evans)10 and the cases that have followed
and elaborated on those principles .11 The test requires the court to consider the
visual, aural and conceptual similarities of the two marks based on the overall
impression and any dominant features. A court should a ssess the impact the marks
would have on the average consumer in the market place who is reasonably we ll-
informed and observant, taking account of the type of goods and how the y are
marketed. The ordinary consumer may not encounter the goods bearing the marks
at the same time and place, and an allowance should be made for the consumer’s
imperfect recollection of the marks. The marks should be viewed side by side as
well as separately. The degree of similarity of the goods should be considered in
relation to the degree of similarity of the marks. The greater the similarity of the
goods the more it may offset some differences in the marks, just as the greater
distinctiveness of the goods may require greater similarity of the marks in order
10 Plascon - Evans Paints (Pty) Ltd v Van Riebeeck Paints (Pty) Ltd 1984 (3) SA 623 (A).
11 Bata v Face Fashions CC and Another 2001 (1) SA 844 (SCA) para 9; Cowbell AG v ICS Holdings Ltd 2001
(3) SA 941 (SCA) para 10.
8
to justify a finding of the likelihood of confusion. It is against this background
that I now turn to determine whether Casadobe’s mark is likely to cause confusion
or deception to the public.
[13] Fratelli contended that CANTO is visually, phonetically and conceptually
confusing ly similar to CANTI , that both words are Italian, CANTI being a plural
of CANTO which means ‘song’ or ‘carol ’ in English. Fratelli contended further
that the one letter difference that is ‘O’ and ‘ I’ does not sufficiently distinguish
CANTO from CANTI. Accordingly, continued the contention, the mark CANTO
is similar to CANTI and the use of the CANTO mark in relation to identical goods
would likely deceive or cause confusion among consumers .
[14] Casadobe argued that even though the first four letters of CANTO and
CANTI are the same, a single letter can make an enormous difference in relation
to trade marks. According to Casadobe, the ‘ I’ and the ‘ O’ at the end of the
respective trade marks, are of fundamental importance for the purpose of
distinction. Adding an ‘ I’ to the end of CANT produces a word that is not found
in the English language and has no meaning to an English speaker. By contrast,
the suffix ‘ O’, when added to the letters CANT, produces a recogni sed English
word with a particular meaning for persons with a literary training and possibly
also brings to mind for many such persons The Cantos of Ezra Pound a long poem
consisting of 116 sections , each of which is a canto.
[15] Casadobe submitted that the two trade marks are not confusingly similar
when compared conceptually for the following reasons. First, that although the
difference in the marks consists of a single letter, tha t difference is significant in
trade marks and the difference in the second syllable also renders the marks
aurally distinguishable. Second, that the Registrar of Trademarks did not flag
9
CANTI as being potentially confusing in the context of the use of CANTO.
Fratelli on the other hand disputes these submissions .
[16] In order to test for deception and confusion, the features of the respective
marks which are dominant must be identified . If they share a dominant feature,
there will ordinarily be a greater likelihood of deception or confusion. Here the
goods to which the marks would be applied are identical. The marks would be
used in relation to sparkling wine s. They would be directed at the same
consumers. The trade channels too, would be the same , that is, outlets, bars and
restaurants. In some instances, the wines might be displayed in close proximity
to each other, in other instance an outlet might carry the one product but not the
other or might display them apart from each other.
[17] Fratelli submitted that it should be borne in mind that consumers will not
always have bottles or menus in front of them when choosing either a wine or
sparkling wine. In a crowded bar or restaurant setting , a consumer would order
beverages from a bar attendant or waiter , the noise would be such that when a
consumer articulates Fratelli’s wine the bar attendant or waiter, as the case may
be, would be deceived or confused into believing that the consumer is asking for
Casadobe’s wine. It submitted further that confusion may arise from the
pronunciation or hearing deficiency on either the part of the consumer or bar
attendant or waiter , of the words CANTO or CANTI. Casa dobe disputed this and
submitted that the a verage consumer’s level of attention varies according to the
category of goods in question. A s wines and sparkling wines are expensive, the
purchase thereof would not be an insignificant purchase and would also not
typically form part of daily domestic shopping. Given their price, submitted
Casadobe , the purchase of wine and sparkling wine would therefore require more
10
discernment on the part of the consumer and would therefore be preceded by a
lengthier examination than would be expected.
[18] I do share this sentiment with Casadobe. Consumers of w ines, including
sparkling wines , are quite discerning . They are likely to exercise circumspection
and a great er degree of care in making a purchase. It is not an overstatement that
many wine drinkers take pride in the product they consume and are able to
distinguish whether it is a spumante or frizzante , its age, as well as its source of
origin. The proper approach is to consider the aural, visual and conceptual aspects
of the marks comprehensivel y rather than in isolation. In my view, there would
therefore be no likelihood of confusion or deception.
[19] As regards the notional market place, I agree with the submission on behalf
of Casadobe that the notional consumer would in no way be confused on
encountering the two marks in the market place as the CANTO mark would be
found on South African wines or MCCs while Fratelli’s CANTI mark would be
found on the imported (Italian ) wine section of the store . However, no evidence
of consumer perception with respect to geographical origin of wine has been
presented. In any event, the comparison in a trade mark i nfringement matter is
between Fratelli’s mark as registered, postulating any normal, fair notional use of
the mark in respect of the goods for which it is registered, and Casadobe’s mark
as it is in fact used. As stated in Plascon -Evans12, the appropriate notional test in
infringement proceedings requires the court to have regard to the normal and
natural use to which Fratelli may put its mark, that is all possible fair and normal
applications of the mark within the ambit of the monopoly created by the terms
of the registration.
12 Plascon -Evans Paints LTD v Van Riebeeck (Pty) Ltd 1984 (3) SA 623 (A) at 641 E -I.
11
[20] It is true that Fratelli’s CANTI trade mark registration covers the identical
goods that Casadobe has admitted selling under the CANTO mark , and it can be
accepted that these wines may be sold in the same retail outlet in South Africa .
CANTI wines would likely be displayed in the foreign or Italian wine section,
while CANTO would be displayed among the local South African wines.
Therefore , the likelihood of deception or confusion is diminished.
[21] As to a visual comparison , both marks make use of a common element, the
word ‘CANT’, which make s for visual similarity. In Swatch AG (Swatch SA) v
Apple Inc13 this Court dealt with comparable marks ‘SWATCH ’ and ‘IWATCH ’
as follows: ‘…the use of the descriptive word, of a commonplace item, serves to place greater
emphasis on the significance of the prefix in each of the marks. It is the ‘I’ in IWATCH that is
the visual differentiator of the mark . So too, the joining of ‘S’ to ‘WATCH’, to make up the
word ‘SWATCH’, provides a visual cue based on the word rather than its component parts.
These features of the mark give rise to a clear visual differentiation.’14
This evaluation of the marks in SWATCH SA is similarly applicable in this case,
with similar results , that CANTO and CANTI are not confusingly similar .
[22] In National Brands Ltd v Blue Lion Manufacturing (Pty) Ltd15 this Court
dealt with trade marks ROMANY CREAMS and ROMANTIC DREAMS. This
Court stated:
‘A word mark , and particularly one that makes use of ordinary language, is not merely a
combination of abstract symbols (at least to the literate observer) but is usually recognisable as
a whole, and for what it conveys … In that respect, in my view, its visual appear ance cannot
be separated altogether from its sense. Where the sense of one word mark differs markedly
from that of another (as in this case), and in particular where the registered mark is well known,
13 Swatch AG (Swatch SA ) v Apple Inc [2021] ZASCA 11; 2021 (3) SA 507 (SCA); 2021 BIP 9 (SCA); [2021]
HIPR 158 (SCA).
14 Ibid para 8.
15 National Brands Ltd v Blue Lion Manufacturing (Pty) Ltd [2001] ZASCA 62; [2001] 4 All SA 235 (A) ; 2001
(3) SA 563 (SCA) .
12
it seems to me that the scope for deception or confusion is reduced, though these are always
matters of degree. In my view, the visual distinction in the words that are in issue in this case,
bearing in mind too that each immediately conjures up a differen t picture, are such that there
is not likely to be deception or confusion as contemplated by s 34(1)(a).’16
In this case, the last letters of CANTO and CANTI determine the different visual
cues, tones and senses.
[23] In Yuppiechef Holdings (Pty) Ltd v Yuppie Gadgets Holdings (Pty) Ltd17,
this Court found that while the word ‘Yuppie’ was common to bot h Yuppiechef
and Yuppie Gadgets, it did not form the dominant feature of the mark. The Court
held that:
‘Those additions (chef and gadgets) are an integral part of the two marks and, in the case of
YUPPIECHEF, the two words have been deliberately combined to form a single composite
word. One cannot simply disregard the additional elements of the two marks. Both visually and
aurally the two marks are different. While ‘yuppie’ is common to both, ‘chef’ and ‘gadgets’
are incapable of being con fused, either when seen or spoken. The one consists of a single
syllable, the other of two. In addition the concepts conj ured up by the words ‘chef’ and
‘gadgets’ are worlds apart.’
[24] Consequently, even though CANTO and CANTI are both five letters long
and have the same first four letters , the ‘I’ and ‘O’ at the end of the marks make
a huge difference. The y determine the different visual cues, tones and senses of
the two ma rks. It is improbable that consumers would be deceived or confus ed.
[25] The finding of the high court that the two marks are closely similar, and
their pronunciation sounds similar, is incorrect. A further finding that the single
letters ‘I’ and ‘O’ which happen to be right at the back of the words does not
sufficiently distinguish the marks from each other is equally incorrect. The
16 Ibid para 10 .
17 Op cit fn 6 para 27 .
13
conclusion that trade marks CANTO and CANTI are visually, aurally and
conceptually similar and that there exist s a likelihood of confusion or deception
among consumers is wrong. Therefore, the appeal must succeed.
Order
1 The appeal is upheld with costs , including the costs of two counsel .
2 The order of the high court is set aside and replaced with the following:
‘The application is dismissed with costs.’
____________________
F E MOKGOHLOA
JUDGE OF APPEAL
14
Appearances
For the appellant: R Michau SC
P Farlam SC
Instructed by: Lovius Block Inc , Bloemfontein
For the respondent: A R Sholto -Douglas SC
L G Kilmartin
Instructed by: Von Siedels Attorneys, Cape Town
Honey Attorneys, Bloemfontein .