See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222651360

Twenty years of agronomic evaluation of wild cocoa trees (Theobroma cacao

L.) from French Guiana

Article in Scientia Horticulturae · August 2007

DOI: 10.1016/j.scienta.2007.05.016

CITATIONS READS

68 263

4 authors, including:

Philippe Lachenaud
French Agricultural Research Centre for International Development
62 PUBLICATIONS 1,375 CITATIONS

SEE PROFILE

All content following this page was uploaded by Philippe Lachenaud on 05 February 2024.

The user has requested enhancement of the downloaded file.

 Scientia Horticulturae 113 (2007) 313–321
www.elsevier.com/locate/scihorti

Review
Twenty years of agronomic evaluation of wild cocoa trees
(Theobroma cacao L.) from French Guiana
Ph. Lachenaud a,*, D. Paulin a, M. Ducamp b, J.-M. Thevenin a
a
CIRAD-Bios, UPR ‘‘Bioagresseurs de Pérennes’’, TA-A31/02, 34398 Montpellier Cedex 5, France
b
CIRAD-Bios, UMR BGPI, TA-A54/K, 34398 Montpellier Cedex 5, France
Received 16 March 2007; received in revised form 25 May 2007; accepted 30 May 2007

Abstract
Almost 500 clones of wild cocoa trees (Theobroma cacao L.) grown from pods collected in 1987 from wild mother-trees in the Camopi and
Tanpok river basins (southeastern French Guiana) have been distributed in around fifteen cocoa producing countries since 1988. The name of those
clones always bears the GU prefix (for ‘‘Guyane’’, i.e. French Guiana). All the germplasm of the same geographical origin present in the CIRAD
collection at Paracou-Combi (Sinnamary, French Guiana), i.e. more than 1600 trees, has been abundantly studied for its morphological
characterization, its agronomic assessment or its genetic diversity. Other assessment work, primarily on resistance to certain diseases, has been
carried out by CIRAD in Montpellier, or in various countries by other organizations.
In order to simplify the choice for breeders faced with a large number of GU clones and wishing to use some of them, an overview is presented
here of the results obtained with this germplasm for various selection criteria such as productivity, the yield:vigour ratio (cropping efficiency), pod
filling, bean size, resistance to pests and diseases, compatibility, sensory qualities, etc. The results obtained for resistance of this material to
witches’ broom disease and black pod rot show a globally high level of resistance, making the GU germplasm a new and major potential source of
resistance to those diseases. The same seems to apply for resistance to mirid damage. The yield levels achieved in French Guiana, along with
cropping efficiency, are noteworthy in some families. The first results acquired reveal a substantial heterosis effect when GU clones are hybridized
with other groups.
Given their distribution in several countries, the known individual qualities and performance of some GU clones present at the Reading
quarantine station (UK) are indicated. A selection of clones that are of interest or promising for incorporation in breeding programmes is provided
for each criterion.
# 2007 Elsevier B.V. All rights reserved.

Keywords: Agronomic traits; Cocoa; French Guiana; Leaf test; Moniliophthora; Phytophthora

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
2. Nomenclature and background. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 314
3. Agronomic performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
3.1. Production and cropping efficiency. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 315
3.2. Pod filling and bean traits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
3.3. Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
4. Resistance to diseases and insects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
4.1. Resistance to black pod rot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 316
4.2. Resistance to witches’ broom disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317
4.3. Resistance to frosty pod rot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
4.4. Resistance to mirids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
5. Technological criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

* Corresponding author. Tel.: +33 4 67 71 56 33; fax: +33 4 67 61 55 81.

E-mail address: philippe.lachenaud@cirad.fr (P. Lachenaud).

0304-4238/$ – see front matter # 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.scienta.2007.05.016
314 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321

6. Value as parents. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318

7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 319
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 320

1. Introduction in zones that are difficult of access (tropical rainforests of South

America), which complicates surveys and the safeguarding of
The cocoa tree (Theobroma cacao L., Malvaceae, formerly collected material. The length of the cocoa tree life cycle
Sterculiaceae) is a tree that originates from the neotropical considerably lengthens breeding programmes: it takes at least 15
rainforests, primarily in the Amazon basin and the Guyana years between planting the first varietal trials and validating the
Plateau. It is a typical cauliflorous plant (not strictly) cross- varieties. It is also necessary to add the time imposed by
pollinated and monoecious, and can reach up to 25 m in height compulsory quarantine between continents to avoid the spread of
in its wild state. The fruit, known as a ‘‘pod’’ is a type of diseases. All that explains why several decades are often seen
indehiscent berry, varying in weight depending on the clones, between the collection of genetic resources in the wild and their
and averaging around 500 g. It contains 30–60 seeds practical use by growers, provided those resources are not quite
surrounded by sweet mucilage. Once fermented and dried, simply lost in the early years, during the safeguarding,
seeds (then known as ‘‘beans’’), constitute the fermented dried characterization and assessment stages. It is thus that, nowadays,
cocoa on which the chocolate industry is based. the most frequently used parents in breeding come from surveys
Cocoa is usually grown under shade, after clearing the forest. (or selections) dating from the 1930s and 1940s. Some major
Planting densities vary from 400 to 2500 (plants/ha) and average surveys have taken place without the collected material being
yields are between 200 and 800 kg of fermented dried cocoa per significantly used: ‘‘there are accessions that derive from
ha (Wood and Lass, 1985). The planting material, which has collecting activities carried out 20, and perhaps 50 years ago,
often undergone little improvement, is generally derived from about which we still know little or nothing of the characteristics
seed. Little use is made of cuttings or budded clones, which are and utility of the genotypes represented in these collections’’
more difficult to grow in the early years of the plantation than (Bartley, 2005). In that respect, surveys that were conducted in
seedlings. The cocoa tree is generally attacked by numerous French Guiana in 1987 can be considered exemplary, given the
insects and diseases: caterpillars, bugs of the mirid family, black limited loss rates observed 20 years after collection and the
pod rot, which is pantropical, witches’ broom and frosty pod rot considerable distribution of clones in producing countries.
in America, and Swollen-Shoot virus in Africa, etc. The International Cacao Genebank Database (ICGD,
Botanists distinguish between two geographical sub-species Wadsworth et al., 2003) lists 730 accessions bearing the GU
of the cocoa tree, T. cacao subsp. cacao and T. cacao subsp. prefix (clones or families of open-pollinated progenies), present
sphaerocarpum, which correspond to the main two groups of in around fifteen countries. However, this number is not
historical cultivars: Criollo and Forastero (Cuatrecasas, 1964). updated, inherent values and associated data are often few and
However, breeders now prefer to consider the existence of several far between, and it could be tricky making a choice for
morpho-geographical groups, on whose hybridization genetic incorporating GU clones in a breeding programme. An
improvement of the plant relies (Lanaud et al., 1999). Good overview of the main results acquired remains to be carried
knowledge of the structuring of the species’ genetic diversity is out in order to identify the potentially most worthwhile
therefore needed to produce high-yielding hybrid varieties. In accessions: it is provided here and the best clones are indicated
that respect, the wild cocoa trees of French Guiana, particularly for the main selection criteria, along with those clones which
the GU accessions discovered following surveys in primary should be avoided in breeding programmes.
forest, prove to be original and, in most studies, irrespective of the
markers, they stand out clearly from the other groups. That 2. Nomenclature and background
particularity affords them great potential interest (Lanaud, 1987;
Lachenaud et al., 1999, 2004; Sounigo et al., 2001). Up to 1989, the GU prefix (for ‘‘Guyane’’, i.e. French
The main selection criteria used in cocoa tree breeding are Guiana) was a code used at the CIRAD Paracou-Combi station
productivity, the yield:vigour ratio (‘‘cropping efficiency’’ or (Sinnamary, French Guiana) for all the cocoa germplasm in the
‘‘yield efficiency’’), bean size, resistance to diseases and collection and could therefore cover quite diverse origins.
insects, along with end-product quality. However, genetic Nevertheless, all the GU clones existing outside French Guiana
control results remain relatively few. originate from the Camopi and Tanpok river basins surveyed in
The cocoa tree, like all other tropical perennial plants, still 1987 (Lachenaud and Sallée, 1993). This GU germplasm only
largely relies on genetic resources of wild origin being available accounts for part of the wild cocoa trees of French Guiana:
for its genetic improvement (especially for disease resistance other accessions come from the basins of the Kérindioutou
criteria). However, given its ecology and biology, incorporating (clones with prefixes KER and B7), Oyapok (OYA, PINA),
such genetic resources into breeding programmes is a lengthy Euleupousing (ELP) and Yaloupi (YAL) rivers (Lachenaud and
and complicated matter. In fact, wild cocoa populations are found Sallée, 1993; Lachenaud et al., 1997).
 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321 315

For any GU clone, the nomenclature adopted is as follows: populations (‘‘demes’’, Hartl and Clark, 1997), with an average
GU xxx–N (or GU xxx/N), where xxx is a three-figure number of 11.2 trees per progeny. Each progeny came from a pod
attributed to the original pod during the surveys, and the N collected from a wild mother-tree and each mother-tree was
suffix is a letter (exceptionally a figure) attributed to a seedling only represented by a single progeny. The trees were monitored
or tree derived from that pod. individually for 10 years, with a view to discovering family and/
During the 1987 surveys, pods were collected from 147 wild or population effects for various agronomic descriptors
mother-trees (and two semi-wild trees), whilst two mother- (Table 1), including the following:
trees were collected in budwood form. A pod from each
mother-tree was preserved in French Guiana, the seedlings - Cumulative yield over 10 years.
obtained were planted in a collection, and then studied - Cropping efficiency at 10 years (ratio of the cumulative yield
individually for 10 years. Other pods, representing only 92 at 10 years to the trunk cross-section, in cm2, 50 cm from the
mother-trees, were sent to Montpellier (France) and, after ground).
quarantine, budwood taken from the seedlings obtained was - Losses due to rot diseases in the field (all types combined).
sent to seven countries (Cameroon, Costa Rica, Ivory Coast,
UK, Philippines, Togo and Trinidad and Tobago) between At the same time, some trees (ortets) were morphologically
1988 and 1990. Each seedling was given an identification characterized for their fruits and beans, from 1994 to 2004
suffix, defining the clone, and it was rare that two countries (Lachenaud and Oliver, 2005a; Lachenaud et al., 2006;
received the same clone. All those clones (around 500) were Lachenaud, 2007), based on IPGRI recommendations (Anon.,
full-sibs (or half-sibs) of the trees studied in French Guiana; in 1981). Of the agronomically useful descriptors studied, this
other words, each wild mother-tree collected in 1987 should article considers:
currently be represented by a family of trees, in the form of
seedlings (in French Guiana) and/or clones (in French Guiana - The average weight of fresh beans per pod (WFB).
and elsewhere). - Pod filling (defined by apparent fertility, AF, which is the
Some GU clones selected in French Guiana from 1992 average number of normal beans in relation to the average
onwards were delivered to the Cocoa Research Unit quarantine number of ovules per ovary).
station in Barbados and to CIRAD in Montpellier. The - The average weight of one fresh bean (AW1B).
quarantine stations in Reading and Barbados also delivered GU
clones to other producing countries (Brazil, Malaysia, India, As the aim of this overview is to provide practical
Indonesia, etc.). It resulted from all those exchanges that it is indications for breeders, only a selection of the set of results
currently difficult to draw up a precise picture of the GU obtained is presented.
material situation, in terms of actual location of the clones, their
evaluation and uses. 3.1. Production and cropping efficiency
Our survey revealed that, in October 2006, all over the
world, the number of surviving clones bearing a ‘‘GU’’ For these two traits, which are fundamental selection
designation amounted to 370. In addition to those clones, criteria, and for which there existed significant ‘‘family’’ effects
mention should also be made of 145 progenies still present in (Table 1), the best clones are presented, along with those to be
the CIRAD collection in French Guiana. avoided, by extrapolating the results from French Guiana
(Table 2). The average production of the wild families was
3. Agronomic performance 342 kg of fermented dried cocoa per ha per year over the first
seven harvests (Lachenaud et al., 2000), whereas the production
The main agronomic results were obtained by individually of a neighbouring control plot of selected hybrids was 712 kg
monitoring 1621 trees in the ‘‘Camopi and Tanpok’’ collection, (Lachenaud et al., 1994). However, the yield levels were
set up at Paracou-Combi in 1988 (Lachenaud et al., 2000; remarkable for certain wild families (equivalent to 1426 kg of
Lachenaud and Oliver, 2001). The collection comprised 144 fermented dried cocoa per ha per year for the first 7 years of
open-pollinated progenies (families) belonging to 11 natural harvests for the best family, GU 285), as was ‘‘cropping

Table 1
Significant effects (with the value of Pr > F in brackets) found after 10 years of individual monitoring in the Paracou-Combi collection (French Guiana)
Criterion ‘‘Population’’ effect ‘‘Family’’ effect ‘‘Individual’’ effect (ortet, clone)
Cumulated production No Yes (<0.0001) –
Cropping efficiency No Yes (<0.0001) –
Resistance to rot diseases in the field Yes (0.0028) – –
WFB Yes (0.0005) – Yes (<0.0001)
Flat beans (%) Yes (<0.0001) – –
Pod filling Yes (<0.0001) – –
Average weight of one bean No – Yes (<0.0001)
Where WFB is the average weight of fresh beans per pod and (–) is the effect not tested.
316 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321

Table 2 Table 3
Best clones, and clones to be avoided, for the potential productivity (PP) and Best clones, and clones to be avoided, for bean criteria (according to data from
cropping efficiency (CE) criteria Paracou-Combi, in French Guiana; Lachenaud and Oliver, 2005a; Lachenaud,
2007 and unpublished data)
Criterion Best clones Clones to be avoided
Best clones Clones to be avoided
PP GU 134-A, B, C GU 114-F, L, P
GU 139-A GU 195-P, V WFB GU 285-A (124) GU 245-A (48)
GU 140-F, S GU 221-C, H, J, V GU 139-A (117) GU 230-A (52)
GU 143-A, B, C GU 251-D, L GU 315-A (117) GU 303-A (54)
GU 144-A, B, C, L GU 261-A, P GU 157-A (115) GU 317-A (57)
GU 280-A GU 314-N GU 252-A (112) GU 311-A (58)
GU 285-A, B, C GU 196-A (59)
GU 286-A, P GU 182-A (59)
GU 303-A, B GU 143-A (60)
GU 325-A
AF GU 238-A (0.80) GU 303-A (0.33)
CE GU 129-A, B GU 114-F, L, P GU 155-A (0.73) GU 245-A (0.40)
GU 134-A, B, C GU 195-P, V GU 280-A (0.73) GU 338-A (0.42)
GU 139-A GU 251-D, L GU 285-A (0.72) GU 182-A (0.43)
GU 140-F, S GU 261-A, P GU 270-A (0.71) GU 317-A (0.44)
GU 285-A, B, C GU 314-N GU 276-A (0.71) GU 116-A (0.44)
GU 286-A, P GU 126-A (0.71) GU 174-A (0.44)
GU 303-A, B GU 266-A (0.71) GU 163-A (0.45)
GU 325-A
AW1B GU 315-A (3.82) GU 311-A (2.20)
For PP, the best clones vary between 804 and 1426 kg of dry cocoa ha 1 year 1 GU 279-B (3.76) GU 254-A (2.22)
for the first seven crops under the conditions at Paracou-Combi (French GU 139-A (3.58) GU 165-A (2.37)
Guiana), and clones to be avoided are below 100 kg. For CE, the best clones GU 285-A (3.53) GU 163-A (2.57)
vary between 0.39 and 0.50, and clones to be avoided are below 0.04 (kg of pods GU 157-A (3.49) GU 344-A (2.58)
cm 2). GU 306-A (2.59)
Where WFB is the average weight of fresh beans per pod (values in brackets, in
g), AF the apparent fertility (ratio of the average number of normal beans to the
efficiency’’ (Lachenaud et al., 2000). For the latter criterion, the average number of ovules per ovary) and AW1B is the average weight of one
average value was 0.16 (as opposed to 0.26 in the controls), fresh bean (in g).
with two wild families outclassing the best control. For these
two criteria, as the best populations proved to be CAM 1 and 3.3. Compatibility
CAM 7, breeders should prefer clones derived from families
GU 126 to 152, 156–161 and 250–286. As compatibility is sometimes considered to be an
agronomic trait, it was also studied for 67 ortets, selected
3.2. Pod filling and bean traits from nine populations (Lachenaud and Oliver, 2005b). They
proved in the majority to be self-incompatible (58%) or very
For the bean traits, which are other important selection slightly self-compatible (34%), i.e. displaying between 15 and
criteria, as they concern the commercial end-product used by 25% of fruit-setting 10 days after manual self-pollination
industry, a list of clones to be preferred and avoided is given (Lachenaud and Oliver, 2005b). Those values classed them
(Table 3). The best clones are those giving the highest values for nonetheless in the ‘‘self-incompatible’’ category according to
WFB, AF and AW1B. The last criterion needs to be over 2.86 g, the definition by Posnette (1945), like Upper Amazon Forastero
the value corresponding to the commercial limit of 1 g for the trees.
average weight of one fermented dried bean. Extrapolation of
the results acquired in French Guiana to other clones has not 4. Resistance to diseases and insects
been possible, since no ‘‘family’’ effect was directly found
(Table 1). However, a significant ‘‘family’’ effect existed for the 4.1. Resistance to black pod rot
‘‘average weight of one pod’’ (Lachenaud et al., 2000), which
was positively correlated with the ‘‘average weight of one This disease, which is the most widespread and probably the
bean’’ (R = 0.76; Lachenaud, 2007). Table 1 reveals significant most damaging, is caused by several species of the genus
‘‘population’’ effects for the average weight of fresh beans per Phytophthora, which attack fruits of all ages, leaves, branches
pod, the flat bean rate and pod filling. Significant differences and the trunk. The species Phytophthora megakarya, which is
were found between ortets for the ‘‘average weight of fresh currently confined to certain African countries, seems to be the
beans per pod’’ and ‘‘average weight of one fresh bean’’ most aggressive in the field (Cilas and Despréaux, 2004).
descriptors (Lachenaud and Oliver, 2005a; Lachenaud, 2007). The only data available for field losses suffered by the GU
Like other Forastero germplasm, in general, GU clones have germplasm are those obtained in French Guiana, in the
relatively small beans and poorly filled pods. However, those collection already mentioned. Only a ‘‘population’’ effect has
results, which were obtained at a single location, need to be been found (Table 1), with the population originating from the
confirmed. Tanpok proving to be more susceptible, whilst the Camopi
 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321 317

populations, Cam 13, 12, 1 and 3 were less affected (Lachenaud resistant or very resistant, and only 6 susceptible (15.0%).
et al., 2000). An analysis of the individual data led to the Seven clones were classed better than the IMC 47 control
selection of 24 high-yielding ortets displaying rot resistance in (Table 4). Only one GU clone has been tested in Cameroon,
the field, of which 18 were tested at CIRAD in Montpellier for GU 255-V, where it has proved to be susceptible.
their resistance to Phytophthora palmivora and P. megakarya,
by early screening on leaf discs. Consequently, the main results 4.2. Resistance to witches’ broom disease
available come from early screening in the laboratory on leaf
discs (Nyassé et al., 1995; Tahi et al., 2000). This disease, caused by the fungus Moniliophthora
perniciosa (previously Crinipellis perniciosa), is only rife in
- In relation to black pod rot caused by P. palmivora, the results South America and on some Caribbean islands (including
of the leaf-tests conducted mostly in Ivory Coast, Ghana and Trinidad). The fungus attacks pods and the tree’s production
France (in the last case with very aggressive strains NS 607 potential (vegetative shoots or flower cushions). The selection
and TRI 1), revealed a generally high level of resistance of resistant planting material has to take into account those
(Anon., 2004; Paulin et al., 2005): out of 57 clones tested, 31 three potential targets, as their resistance to the disease is only
(54.4%) proved to be resistant or very resistant, 17 moderately moderately correlated (Pires et al., 1999; Thevenin et al., 2005).
resistant (29.8%) and 6 susceptible or very susceptible In French Guiana, both in situ and ex situ, GU germplasm
(10.5%). The ambiguous cases (clones scored differently proves to be remarkably resistant to the local strains of the
depending on the country) were rare (3, i.e. 5.3%), like the disease, to such a point that no susceptibility scoring is possible
example of clone GU 255-V, which varied from susceptible in the collection.
(in Venezuela) to very resistant (in Ivory Coast), undoubtedly In Trinidad, the results on pods obtained in the field on nine
depending on the aggressiveness of the strains used (Anon., clones showed that eight were resistant and one moderately
2004). resistant (Table 4). On twigs, in Ecuador and Trinidad (Anon.,
- In relation to black pod rot caused by P. megakarya, the leaf 2004) and in Brazil (Pires, 2003; Anon., 2004), out of 28 clones
tests conducted in Montpellier (Paulin et al., 2006), with a tested, 22 were resistant (78.6%), 4 susceptible (14.3%) and 2
very aggressive strain (NS 269), revealed generally high displayed contradictory results. In greenhouse tests in Trinidad
resistance: out of 40 clones tested, 28 (70.0%) proved to be (test by spraying or using agar droplets), out of nine clones,

Table 4
Best clones (VR, very resistant and/or R, resistant) and clones to be avoided (S, susceptible; VS, very susceptible) in controlling three diseases and a group of cocoa
tree pests
Best clones Clones to be avoided
P. palmivora (leaf test)
Strain TRI 1 GU 123-V (VR)
GU 226-V, GU 195-V
GU 269-V, GU 241-V (R)
Strain NS 607 GU 134-B, GU 254-A (R) GU 138-A (VS)
GU 230-B, GU 268-A
GU 134-C, GU 301-A (S)
P. megakarya (leaf test) (strain NS 269) GU 123-V, GU 125-C GU 129-B, GU 134-A
GU 195-V, GU 226-V GU 138-A, GU 265-V
GU 269-V (VR) GU 312-V, GU 329-V (S)
M. perniciosa (in the field on twigs and/or pods) GU 123-C, GU 125-C GU 168-H, GU 259-C
GU 151-F, GU 171-C GU 296-H, GU 341-H
GU 175-V, GU 195-P (S)
GU 219-F, GU 221-C
GU 221-H, GU 241-P
GU 261-P, GU 265-P
GU 277-G, GU 286-P
GU 300-P, GU 305-P
GU 307-F, GU 307-V
GU 310-P, GU 335-P
GU 353-L (R)
Mirids GU 183-K, GU 191-B GU 125-K, GU 158-G
GU 207-G, GU 207-K (S)
GU 226-R, GU 265-K
GU 275-A, GU 298-B
GU 322-B, GU 333-K
GU 337-F, GU 343-F (VR)
The clones for which ambiguity persists are not included.
318 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321

only one (GU 255-V) proved to be susceptible (whereas it was parameters, such as acidity, astringency, floral or fruity tastes,
scored resistant in the field in the three countries mentioned), aroma intensity, aftertaste, cocoa flavour, etc. They concern the
five were resistant and three moderately resistant (Anon., 2003, end-products used by industry and consumed.
2004). Pires (2003) considered that the clones of the GU series The sensory qualities of the GU material have yet to be
were, on the whole, a new group of substantial interest, studied extensively, though it has been done for some
displaying resistance factors that were distinct from those of populations, using fermented dried cocoa primarily produced
SCA 6 (or of the Trinitarios and Lower Amazons), and in French Guiana (Assemat et al., 2005) and for one clone, GU
therefore potentially very worthwhile in crosses with those 175-P, in Ecuador (Anon., 2006).
materials. In the first work, with samples of dry cocoa produced in
French Guiana and Ivory Coast, the overall aroma and
4.3. Resistance to frosty pod rot ‘‘flavour’’ intensity of chocolates produced from GU germ-
plasm has proved to be statistically better than that of the West
This disease, caused by Moniliophthora roreri, is currently African Amelonado which, when it comes from Ghana, is the
confined to certain American countries, but its recent and rapid industrial reference for those criteria (Assemat et al., 2005).
invasion of Guatemala and Mexico, along with its current The caffeine rate was also higher than that of the African
spread towards Bolivia and Brazil, make this a potentially very Amelonado control. For the other criteria studied by Assemat
dangerous disease for world cocoa growing. A thick layer of et al. (2005), especially the butter rate, no significant difference
white mycelium develops on the surface of the pods, containing was found with the controls (various Amelonados and
billions of spores that can easily be disseminated by wind, Ecuadorian cocoas).
water, insects or humans. A major programme to search for In the second work, clone GU 175-P, tested using fermented
resistant cultivars has been under way for around 12 years in dried cocoa produced in Ecuador, had ‘‘a rich cocoa aroma with
Costa Rica; however, sources of resistance to this disease are raisin notes’’ (Anon., 2006). Its average bean weight was 1.09 g
few in number and need to be enhanced. (as opposed to 0.63 g for clone SCA 6, a reference for small
In Costa Rica (applying a methodology described in beans) and its fat content was quite high (54.2%, for an average
Phillips-Mora et al., 2005), out of seven GU clones tested of 53.7 over the 15 samples of various morphogeographical
(W. Phillips-Mora, personal communication), four proved to be groups of cocoa trees).
susceptible (GU 168-N, GU 200-L, GU 237-N, GU 269-H) and
three ‘‘moderately susceptible’’ (GU 151-N, GU 154-L and GU 6. Value as parents
207-N).
Data in Tables 2–5 concern clonal values of the GU
4.4. Resistance to mirids germplasm (ortets, clones); they could be useful for choosing
clones, but are particularly so for selecting parents, since the
Mirids (often called capsids) are a group of bugs that are rife cocoa tree is primarily used in the form of inter-group hybrid
in western and central Africa, whose most harmful species varieties.
(Sahlbergella singularis and Distantiella theobromae) cause GU clones have only undergone limited testing as parents in
damage to branches and pods, but can also kill trees. In cocoa breeding programmes, with generally quite recent trials,
addition, the lesions caused by mirids provide access for but the following are worth mentioning:
various fungi that can obstruct vessels (Wood and Lass, 1985;
Collingwood, 1977). Some GU clones have been tested against - In Ivory Coast (CNRA), two clones were involved in four
these insects in Ivory Coast, Cameroon and Ghana. ‘‘single pair’’ type crosses planted in 1991 (Lachenaud et al.,
In the CNRA collection in Ivory Coast, the scores for old and 2001) and three (involved in 11 crosses) in a 6  3 incomplete
recent damage (N’guessan et al., 2007) showed the group of GU factorial design at two locations in 2001, after leaf-testing of
clones to be one of the most resistant groups to those insects (the the families for resistance to P. palmivora (Anon., 2004). As a
most resistant for old damage and among the three most resistant result of the first trial, a progeny with parent GU 133-1
for recent damage). Twelve clones showed no damage after a (classed resistant to P. palmivora) has been shortlisted for
dozen years in the field (Table 4). In Cameroon, in budwood multi-site confirmation trials, due to its high productivity,
gardens, with artificial infestation by S. singularis larvae, GU excellent cropping efficiency and very low losses due to rot
255-V, the only GU clone tested, proved to be resistant (Anon., diseases (Lachenaud et al., 2001). Some of the results
2004; Babin et al., 2004; Babin, personal communication). This acquired seem to show good transmission of resistance to P.
result is confirmed in Ghana (Adu-Acheampong et al., 2007) palmivora: for instance, the progenies of clone GU 123-B
were this clone is among 11 tolerant to mirid attacks. (with parents P7, IFC 1 and IFC 11), which was classed highly
resistant in the leaf test, were also classed highly resistant
5. Technological criteria (Anon., 2004). In the field, the first harvests (Anon., 2006 and
Tahi, personal communication) revealed an excellent perfor-
The technological criteria comprise traits such as bean size mance for yield and resistance to rot diseases for hybrids with
(already mentioned in Section 3.2; Table 3), chemical a GU parent (in this case GU 175-A, GU 123-B, GU 284-A)
composition (fat content, alkaloids) and various sensory planted in comparative trials with other families in 2001.
 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321 319

Table 5
Known characteristics of GU clones in quarantine at the University of Reading (2006)
Name Population Production CE Rot in Phytophthora Phytophthora WB WB WB Mirids
field palmivora test megakarya test test pods twigs
GU 114-P TAN VL VL S R MR R-S
GU 123-V TAN – – VR VR
GU 125-C CAM 6 L L VR R
GU 133-C CAM 7 L L VR
GU 144-C CAM 7 H H
GU 168-H CAM 9 L L S
GU 171-C CAM 9 L L R
GU 175-P CAM 9 L L MR R R
GU 183-G CAM 9 L L R R VR
GU 195-V CAM 9 VL VL R VR
GU 207-H CAM 12 L L VR R
GU 219-F CAM 3 L L MR MR R R
GU 221-C CAM 3 VL L R
GU 226-V CAM 3 A L R VR
GU 241-P CAM 3 L L MR R R
GU 241-V CAM 3 L L R R
GU 259-C CAM 1 L L S
GU 261-P CAM 1 VL VL R
GU 263-V CAM 1 A A R R
GU 265-P CAM 1 L L R
GU 269-V CAM 1 L L R VR
GU 277-G CAM 1 H H R R VR
GU 296-H CAM 13 L L S
GU 307-F CAM 9 L L S R R
GU 310-P CAM 9 L A S R R
Where CE is the cropping efficiency, WB witches’ broom, MR moderately resistant, R resistant, VR very resistant and S is susceptible. Production, in kg of dry cocoa
ha 1 year 1: H, high (>750); A, average (400–750); L, low (100–399); VL, very low (<100); CE, in kg of pods cm 2. H, high (>0.30); A, average (0.20–0.29); L,
low (0.04–0.19); VL, very low (<0.04).

- In Trinidad (CRU), in a pre-breeding programme for other clones belonging to other groups (EET 103, ICS 95 and
resistance to black pod rot, three clones have been involved GF 23). The Specific Combining Ability was significant and
in a 6  6 half diallel design (Ducamp, 1994), initially carried the best crosses were those involving an Upper Amazon
out for nursery tests, then later planted for certain families. In parent and a GU parent, whilst the least good were all crosses
the nursery, 29 progenies, along with buddings of the parents, between GU clones. Those results need to be confirmed for
were assessed for their degree of resistance to P. palmivora, other agronomic criteria but, like those obtained in Ivory
between 1995 and 2000, using the leaf disc inoculation Coast, they seem to indicate the considerable merits of the GU
method (Nyassé et al., 1995). The three Guianan clones used group in hybridization.
as parents (GU 175-P, GU 286-P and GU353-L) displayed
moderate resistance levels (from 2.65 to 2.88, when Crosses involving GU clones have also been planted recently
symptoms were scored from 0 to 5). By crossing with clones in other countries. In America, a cross of this type (GU 154-
SCA 6, IMC 67 and ICS 1, the progenies had lower resistance L  ICS 43) is involved in a regional trial planted between
levels than the parental average, but not significantly so 2004 and 2006 (Costa Rica, Peru, Trinidad, Ecuador, Brazil),
(r = 0.65, pr = 0.12). That germplasm was planted out and and in Ghana (CRIG), 14 GU clones were crossed with Upper-
thereby subjected to natural black pod rot and witches’ broom Amazon clones (17 crosses) and planted in 2002 (Anon., 2004).
pressure. Three years of observations on branches were used Only a few preliminary results are available: for instance, in
to quantify the percentage of buds infected by M. perniciosa the Costa Rican trial, cross GU 154-L  ICS 43 is classed first
(unpublished data): clones GU 175-P and GU 286-P proved to (out of 13 crosses) for trunk circumference and crown height.
be very resistant (0.44 and 0.00%, respectively) and GU 353- The particular vigour of the cross is attributed to the GU parent
L susceptible (4.61%). (Anon., 2006), a particularity that seems to be confirmed in
- In French Guiana (CIRAD), a genetic parameters study was Brazil (Monteverde-Penso et al., 2005).
conducted. Seven clones were involved in 27 crosses, in a
5  8 incomplete factorial trial, planted in 1997–1998. For 7. Conclusion
the juvenile vigour criterion (increase in trunk section 15 cm
from the ground between 1 and 2 years in the field), some The GU germplasm primarily surveyed in 1987 is still well-
unpublished results indicated that GU clones had a lower represented 20 years after it was collected: 94% of the wild
General Combining Ability than the two Upper Amazon mother-trees still have representatives in collections, which is
clones tested (IMC 67 and SCA 6), but similar to that of the quite an exceptional result. Bartley (2005) pointed out that the
320 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321

Brazilian Genetic Resources Programme had lost more than Acknowledgements

50% of the mother-trees collected since 1976. The good results
of the CIRAD programme can be partly explained by the fact Our thanks to P. Biggins for translating the text and to all the
that virtually only pods were taken from the mother-trees (the researchers who agreed to respond to our survey.
budding success rate is always lower) and that the material was
very rapidly distributed to several countries. It currently exists References
in seventeen producing countries.
Twenty years after its collection in primary forests, this Adu-Acheampong, R., Padi, B., Ackonor, J.B., Adu-Ampomah, Y., Opuku, I.Y.,
material is still being assessed, after largely being character- 2007. Field performance of some local and international clones of cocoa
against infestation by mirids. In: Eskes, A.B., Efron, Y. (Eds.), Global
ized in French Guiana, and has already been incorporated in Approaches to Cocoa Germplasm Utilization and Conservation. CFC
some national breeding programmes. The results obtained on Technical Paper No. 50, pp. 187–189.
the resistance to witches’ broom disease and black pod rot Anon., 1981. Genetic Resources of Cocoa. Report ‘‘IBPGR Working Group on
have revealed a high level of general resistance, which Genetic Resources of Cocoa’’. IBPGR Secretariat, Roma, 25 pp.
confirms the observations carried out in situ during the Anon., 2003. Evaluation of Cocoa Germplasm for Resistance to Witches’
Broom Disease. Report to ACRI. Cocoa Research Unit, University of
surveys and ex situ in Guianan collections. The group of GU the West Indies, 10 pp.
clones therefore proves to be a major potential source of Anon., 2004. CFC/ICCO/IPGRI Project on: Cocoa Germplasm Utilization and
resistance to those diseases. The situation seems to be similar Conservation: A Global Approach. Final Individual Reports April 1998–
for resistance to mirid damage, but fewer results are September 2003, vol. 1 and 2.
Anon., 2006. CFC/ICCO/IPGRI Project on: Cocoa Productivity and Quality
available.
Improvement, A Participatory Approach. Appendix to the General Progress
The production levels reached in French Guiana, along with Report Year 2. Individual Institute Reports, 1 June 2005–31 May 2006.
the cropping efficiency, are noteworthy for some families. The Assemat, S., Lachenaud, Ph., Ribeyre, F., Davrieux, F., Pradon, J.-L., Cros, E.,
best progenies have produced yields approaching 3000 kg of 2005. Bean quality traits and sensory evaluation of wild Guianan cocoa
dry cocoa per hectare once in full production (Lachenaud et al., populations (Theobroma cacao L.). Genet. Resour. Crop Evol. 52, 911–917.
2000). The first results obtained in Ivory Coast showed that GU Babin, R., Sounigo, O., Dibog, L., Nyassé, S., 2004. Field tests for antixenosis
and tolerance of cocoa towards mirids. Ingenic Newsletter 9, 45–50.
clones seem to reveal a high heterosis effect for yield in Bartley, B.G.D., 2005. The Genetic Diversity of Cacao and its Utilization. CABI
hybridization with other widely used groups. Publishing, UK and USA, 341 pp.
Bean-related criteria have shown that, in general, GU clones Cilas, C., Despréaux, D., 2004. Improvement of Cocoa Tree Resistance to
have relatively small beans and poorly filled pods, which are Phytophthora Diseases. Cirad. Collection Repères, 176 pp.
common defects of Forastero cocoa trees. However, there exists Collingwood, C.A., 1977. African mirids. In: Lavabre, E.M. (Ed.), Les Mirides
du Cacaoyer, IFCC et ed. G.-P. Maisonneuve et Larose, Paris, 367 pp.
substantial variability and some worthwhile clones can be Cuatrecasas, J., 1964. Cacao and its allies. A Taxonomic Revision of the Genus
selected and recommended. The sensory value, about which Theobroma. Contributions from the United States National Herbarium, vol.
little is known yet, seems nonetheless interesting, with a strong 35, no. 6. Smithsonian Institution, Washington, DC.
cocoa flavour. Ducamp, M., 1994. Evaluation of Cacao Genetic Resources for Resistance to
Black Pod and Witches’ Broom Diseases in Trinidad. The University of the
This overview, along with the results already published on
West Indies, Cocoa Research Unit (Report for 1994), 49–52.
the spatial structuring of the genetic diversity found in GU Hartl, D.L., Clark, A.G., 1997. Principles of Population Genetics, third ed.
material (Lachenaud et al., 2004), should therefore provide Sinauer Ass. Inc., Sunderland, MA, USA.
breeders with a relatively easy choice of parents to be Lachenaud, Ph., 2007. Fruit trait variability in wild cocoa trees (Theobroma
incorporated into their breeding programmes. GU clones also cacao L.) from the Camopi and Tanpok basins in French Guiana. Acta Bot.
possess the interesting property of being virtually homozygous, Gallica 154 (1), 117–128.
Lachenaud, Ph., Sallée, B., 1993. Les cacaoyers spontanés de Guyane. Loca-
with 6–11% heterozygosity being observed with biochemical lisation, écologie, morphologie. Café Cacao Thé (Paris) 37, 101–114.
markers (Lachenaud et al., 2004). Lachenaud, Ph., Oliver, G., 2001. Variability in various agronomic traits of wild
For the whole set of criteria – yield, cropping efficiency, cocoa trees (Theobroma cacao L.) from the Camopi and Tanpok basins
bean size, average weight of fresh beans per pod, apparent (French Guiana). PGR Newsletter 128, 35–40.
Lachenaud, Ph., Oliver, G., 2005a. Variability and selection for morphological
fertility, resistance to rot diseases in the field confirmed by the
bean traits in wild cocoa trees (Theobroma cacao L.) from French Guiana.
leaf test – three clones selected in French Guiana and Genet. Resour. Crop Evol. 52, 225–231.
characterized can be recommended: GU 285-A, GU 134-B Lachenaud, Ph., Oliver, G., 2005b. Compatibility and duration of pod matura-
and GU 139-A. These three clones, used as parents in tion in Guianan wild cocoa trees—preliminary results. Ingenic Newsletter
breeding programmes, could provide significant genetic 10, 20–22.
progress. In addition, some clones in quarantine in Reading Lachenaud, Ph., Clément, D., Sallée, B., Bastide, Ph., 1994. Le comportement
en Guyane de cacaoyers sélectionnés en Côte d’Ivoire. Café Cacao Thé 38,
(UK), and therefore available and easily accessible, display 91–102.
worthwhile resistance capacities (Table 5). Five of them are Lachenaud, Ph., Mooleedhar, V., Couturier, C., 1997. Wild cocoa trees in
also included in the CFC/ICCO/IPGRI collection (Sounigo French Guiana. New Surveys. Plant. Rech. Dév. 4, 25–32.
et al., 2007): among them, GU 175-P, GU 241-P and GU 261- Lachenaud, Ph., Bonnot, F., Oliver, G., 1999. Use of floral descriptors to study
P are resistant to witches’ broom disease and, in the first two variability in wild cocoa trees (Theobroma cacao L.) in French Guiana.
Genet. Resour. Crop Evol. 46, 491–501.
cases, display a degree of resistance to P. palmivora. GU 133- Lachenaud, Ph., Oliver, G., Letourmy, Ph., 2000. Agronomic assessment of wild
C (and in most cases GU 255-V) displays high resistance to P. cocoa trees (Theobroma cacao L.) from the Camopi and Tanpok basins
palmivora. (French Guiana). PGR Newsletter 124, 1–6.
 P. Lachenaud et al. / Scientia Horticulturae 113 (2007) 313–321 321

Lachenaud, Ph., Eskes, A.B., N’goran, J.A.K., Clément, D., Kébé, I., Tahi, M., bian isolates of Moniliophthora roreri from four pathogen genetic groups.
Cilas, C., 2001. Premier cycle de sélection récurrente en Côte d’Ivoire et Plant Pathol. 54, 483–490.
choix des géniteurs du second cycle. In: Proceedings of the 13th Interna- Pires, J.L., 2003. Avaliação quantitativo e molecular de germoplasma para o
tional Conference on Cocoa Research, Kota Kinabalu, Sabah, Malaysia, 9– melhoramento do cacaueiro com enfase na produtividade dos frutos e
14 October 2000, pp. 11–22. rêsistencia. Ph.D. Thesis. Universidad Federal de Viçosa, MG, Brazil,
Lachenaud, Ph., Sounigo, O., Oliver, G., 2004. Genetic structure of Guianan 242 pp.
wild cocoa (Theobroma cacao L.) described using isozyme electrophoresis. Pires, J.L., Monteiro, W.R., Luz, E.D.M.N., Silva, S.D.V.M., Pinto, L.R.M.,
PGR Newsletter 139, 24–30. Figueira, A., Gramacho, K.P., Lopes, U.V., Bevilaquia Albuquerque, P.S.,
Lachenaud, Ph., Oliver, G., Bastide, Ph., Paulin, D., 2006. Le remplissage des Yamada, M.M., Ahnert, D.E., Brugnerotto, M.I.B., 1999. Cocoa breeding
cabosses des cacaoyers spontanés de Guyane (Theobroma cacao L.). Acta for witches’ broom resistance at CEPEC, Bahia, Brazil. In: Proceedings of
Bot. Gallica 153, 105–114. the International Workshop on the Contribution of Disease Resistance to
Lanaud, C., 1987. Nouvelles données sur la biologie du cacaoyer (Theobroma Cocoa Variety Improvement, Salvador, Bahia, Brazil, 24–26 November
cacao L.). Thèse de doctorat d’état. Université de Paris XI, Orsay, France, 1996, INGENIC, Reading, UK, pp. 91–101.
262 pp. Posnette, A.F., 1945. Incompatibility in Amazon cacao. Trop. Agric. (Trinidad)
Lanaud, C., Motamayor, J.C., Sounigo, O., 1999. Le cacaoyer. In: Hamon, P., 22, 184–190.
Seguin, M., et al. (Eds.), Diversité génétique des Plantes Tropicales Sounigo, O., Bekele, F., Iwaro, D., Thevenin, J.-M., Bidaisse, G., Umaharan, R.,
Cultivées. Collection Repères. CIRAD, 387 pp. Sankhar, A., Sukha, D., Boccara, M., Butler, D., Eskes, A.B., 2007.
Monteverde-Penso, E.J., Phillips, W., Machado, R., Brown, J.S., Schnell, R., Description of cocoa clones proposed for the ‘‘CFC/ICCO/IPGRI Project
Motamayor, J.C., 2005. Can we select for shorter cocoa plants? Ingenic Collection’’. In: Eskes, A.B., Efron, Y. (Eds.), Global Approaches to Cocoa
Newsletter 10, 16–19. Germplasm Utilization and Conservation. CFC Technical Paper No. 50, pp.
N’guessan, K.F., Eskes, A.B., Lachenaud, Ph., 2007. Mirid resistance studies in 67–81.
Côte d’Ivoire: field observations on recent and cumulative damage. In: Sounigo, O., Ramdahin, S., Christopher, Y., 2001. Etude de la diversité
Eskes, A.B., Efron, Y. (Eds.), Global Approaches to Cocoa Germplasm génétique du cacaoyer à l’aide des RAPD. In: Proceedings of the 13th
Utilization and Conservation. CFC Technical Paper No. 50, pp. 170–181. International Conference on Cocoa Research, Kota Kinabalu, 9–14 October
Nyassé, S., Cilas, C., Hérail, C., Blaha, G., 1995. Leaf inoculation as an early 2000, Malaysia, pp. 119–126.
screening test for cacao (Theobroma cacao L.) resistance to Phytophthora Tahi, G.M., Kébé, B.I., Eskes, A.B., Ouattara, S., Sangaré, A., Mondeil, F.,
black pod disease. Crop Prot. 14, 657–663. 2000. Rapid screening of cacao genotypes for field resistance to Phy-
Paulin, D., Ducamp, M., Vezian-Bonnemayre, K., Eskes, A.B., 2005. Evalua- tophthora palmivora using leaves, twigs and roots. Eur. J. Plant Pathol. 106,
tion par le test d’inoculation sur feuille du niveau de résistance de trente 87–94.
clones vis-a-vis de plusieurs espèces de phytophthora. In: Proceedings of Thevenin, J.-M., Umaharan, R., Surujdeo-maharaj, S., Latchman, B., Cilas,
the 14th International Conference on Cocoa Research, vol. 1, Accra, Ghana, C., Butler, D.R., 2005. Relationships between black pod and
13–18 October 2003, pp. 85–93. witches’ broom diseases in Theobroma cacao. Phytopathology 95,
Paulin, D., Ducamp, M., Lachenaud, Ph., 2006. Evaluation des cacaoyers 1301–1307.
spontanés de Guyane française pour leur résistance à Phytophthora Wadsworth, R.M., Ford, C.S., Turnbull, C.J., Hadley, P., 2003. International
megakarya. In: Proceedings of the 15th International Conference on Cocoa Cocoa Germplasm Database, v.5.2. Euronext.liffe/University of Reading,
Research, San José, Costa Rica, 9–14 October (due out). UK.
Phillips-Mora, W., Castillo, J., Krauss, U., Rodriguez, E., Wilkinson, M.J., Wood, G.A.R., Lass, R.A., 1985. Cocoa, Tropical Agriculture Series, fourth ed.
2005. Evaluation of cacao (Theobroma cacao) clones against seven Colom- Longman, New York, 621 pp.

View publication stats