Carol J.

Lovatt
Professor of Plant Physiology
Department of Botany and Plant Sciences
University of California, Riverside

Alternate Bearing Of ‘Hass’ Avocado

A summary of basic information to assist growers in managing their orchards
Concepts reviewed in this summary –
• The problem of alternate bearing.
• Fruit influence avocado tree phenology and return bloom and yield.
• Summer vegetative shoot growth is key to a good return bloom and yield.
• The setting ON crop of fruit inhibits summer vegetative shoot growth.
• The setting ON crop of fruit has a localized and whole tree effect on return bloom.
• The number of shoots that do not set fruit during the ON-crop year is important.
• The effect of the ON crop on return bloom is cumulative.
• Year 1 mature avocado fruit do not influence flower and fruit abscission in Year 2.
• Strategies growers can use now to mitigate alternate bearing.
• We are all in this together, share results!
The problem of alternate bearing
Alternate bearing (also called biennial or uneven bearing) is the tendency of a fruit tree to produce a
heavy, high yield “ON crop” (ON tree, ON year) followed by a light, low yield (or no crop) “OFF
crop” (OFF tree, OFF year). The ON year is initiated with an intense ON bloom that sets the ON crop,
whereas the OFF year is initiated with a sparse OFF bloom that sets the OFF crop. The phenomenon is
widespread, occurring in both deciduous and evergreen trees under commercial production and in
nature, where it is called “masting” (Monselise and Goldschmidt, 1982). Alternate bearing can occur
over a large geographical area, an entire grove, an individual tree, part of a tree or even for one branch
(Monselise and Goldschmidt, 1982).
Alternate bearing has a severe negative economic impact on the commercial avocado industry. The
ON crop is characterized by a large number of small-sized fruit of reduced commercial value at
harvest. For the OFF crop, fruit size is large but there are too few fruit to provide the grower with a
good income. Annual yield and crop value data for the California avocado industry summarized in
Table 1 reflect the reduced cash value of the ON-crop fruit and the limited revenue generated by the
OFF crop. The data in Table 1 document that alternate bearing ‘Hass’ avocado orchards suffer a
measurable loss in profitability in both the OFF- and ON-crop years. In addition, the reduced amount
of fruit in the OFF year can result in a loss of market share and threaten the sustainability of the fresh
fruit industry. Years with low fruit volume can compromise the development and sustainability of
value-added product industries. Furthermore, alternation in crop load complicates orchard
management.
Keep in mind that the impact of alternate bearing on individual growers is more dramatic than the
California avocado industry averages reveal. The alternate bearing index (ABI) reflects the severity of
alternate bearing in an orchard. In its simplest form, ABI can be calculated by subtracting the yield for
the OFF-crop year from the yield for the ON-crop year and dividing the difference by the sum of the
ON and OFF yields. When ABI is 1.0, alternate bearing is 100% (i.e., crop one year, no crop the other
year) and when ABI is 0, there is no alternate bearing (Pearce and Dobersek-Urbanc, 1967). Typical
ABI values for ‘Hass’ avocado orchards in California are between 0.57 and 0.97, which indicates that
yield is 60% to 100% less in the OFF-crop year compared to the ON-crop year.
The problem of alternate bearing reoccurs because the major factors initiating alternate bearing are
recurring climatic events. Alternate bearing is typically initiated by freeze, low or high temperatures,
or water-deficit stress during bloom or fruit set, which result in low flower number or cause excessive
flower and/or fruit drop. This results in an OFF crop, which is typically followed by an ON crop the
next year, depending on how long it takes the trees to recover. Alternate bearing can also be triggered
by under fertilization and irrigation, resulting in excessive flower and fruit abscission, or by over
pruning, resulting in low flower or fruit number and excessive vegetative shoot growth. These
conditions result in an OFF crop that will be followed by an ON crop. Conversely, optimal conditions
for bloom and fruit set such that natural fruit thinning fails to take place result in an ON crop followed
by an OFF crop. Once initiated, alternating ON/OFF cycles are perpetuated by the effect of fruit
number (crop load) on factors within the tree that ultimately impact floral intensity the following year.
Thus, an ON bloom sets and matures an ON crop that causes a low floral intensity OFF bloom and
OFF crop the following year, whereas the OFF bloom sets and matures an OFF crop that results in an
intense ON bloom the next spring. Since climate is a major factor initiating alternate bearing, the need
for a solution to the problem of alternate bearing is frequent, but the need to culturally manage
alternate bearing is almost annual.
Fruit influence avocado tree phenology and return bloom and yield
In recent research, the effects of the OFF crop, the ON crop and removing the entire ON crop on
‘Hass’ avocado tree phenology, vegetative shoot growth, floral development and the next year’s yield
under California growing conditions were quantified. The results of this research provide insight into
solutions to alternate bearing that can be used now. In addition the results identified the best time for
taking action and established the consequences of delaying action or doing nothing.
For the sake of this discussion, Year 1 starts with bloom. The ON crop (ON year) is initiated with an
intense ON bloom, followed by the setting of the ON crop of fruit. We will follow the Year 1 ON crop
from fruit set through development to harvest to discuss how and when the ON crop of fruit at each
stage of development impacts avocado tree phenology and return bloom and yield the following year
(Year 2). To facilitate comparison of the effects of ON and OFF avocado crops, we will similarly
discuss the effects of the OFF crop of fruit on avocado phenology starting with the Year 1 light OFF
bloom, which sets the OFF crop.
Summer vegetative shoot growth is key to a good return bloom and yield
OFF-bloom/OFF-crop Year. When the avocado tree sets and develops an OFF crop, a significant
number of vegetative shoots develop during the summer and to a lesser degree during the fall of Year
1 (Fig. 1). The following spring, it is the Year 1 summer shoots that contribute the greatest number of
floral shoots (inflorescences) to bloom, resulting in the ON bloom and ON crop in Year 2. For OFF-
crop ‘Hass’ avocado trees in California, ~75% of the Year 2 floral shoots are produced by the Year 1
summer vegetative shoots, ~15% by the Year 1 spring vegetative shoots and ~10% by Year 1 fall
vegetative shoots. Note that Year 1 summer shoots produce ~80% of the determinate floral shoots that
contribute to making Year 2 an ON bloom (Lovatt, 2004, 2005, 2006). Determinate floral shoots,
which are more productive than indeterminate floral shoots, are produced by the apical and axillary
(lateral buds) borne in the axil of the leaves along the length of summer vegetative shoots. Fall shoots
that flower produce only a single terminal indeterminate floral shoot (Salazar-García and Lovatt,
1998). Thus, summer vegetative shoot growth is key to a good return bloom; the more summer
vegetative shoots, the greater the return bloom and yield.
Figure 1. An OFF-crop year starts with an OFF bloom that sets an OFF crop or is initiated by a climatic or culture event
that reduces flower or fruit number during the bloom or fruit set periods. The low number of young developing fruit in Year 1
results in a significant amount of summer and fall vegetative shoot growth that contributes floral shoots to the ON bloom in
Year 2, ~75% and ~10% by summer and fall shoots, respectively. Summer vegetative shoots also contribute the majority of
determinate floral shoots, which are present in high numbers only in an ON bloom. Note that the mature fruit of the OFF
crop are fewer in number but large in size.

The setting ON crop of fruit inhibits summer vegetative shoot growth

ON-bloom/ON-crop Year. During Year 1, the ON crop of young developing fruit inhibits summer and
fall vegetative shoot growth, thereby reducing the number of buds that can produce floral (and
vegetative) shoots the next spring, resulting in an OFF bloom (Year 2) (Fig. 2). Note that due to the
inhibition of summer vegetative shoot growth during the ON-crop year, there are few to no
determinate floral shoots at spring bloom in Year 2 and the majority of the floral shoots that do
develop are produced by the Year 1 spring shoots.

Figure 2. An ON-crop year starts with an ON bloom that sets an ON crop or is initiated by climatic or culture events that are
optimal for flowering and fruit set such that natural fruit thinning fails to take place. The large number of young developing
fruit inhibits summer and fall vegetative shoot growth. The loss of summer and fall shoots means there are fewer sites that
can bear floral shoots in spring of Year 2, especially determinate floral shoots, which develop predominantly on summer
shoots. The result is an OFF bloom in Year 2 that is produced predominantly by the Year 1 spring shoots. Note that there are
many smaller mature fruit in the ON crop.

The importance of summer and fall shoots to the floral intensity of bloom in Year 2 was confirmed by
removing all summer and fall shoots from OFF-crop ‘Hass’ avocado trees during Year 1 and
comparing the return bloom and return yield of these trees to untreated OFF-crop control trees. At
return bloom the following year, untreated OFF-crop control trees flowered profusely, whereas OFF-
crop trees with summer and fall shoots removed had a weak bloom and low return yield (2 kg/tree in
Year 2) (Lopez-Jimenez and Lovatt summarized in Lovatt, 2005). From these data it is clear that
summer vegetative shoot number is key to a good return bloom and yield and that the ON crop of fruit
reduces summer vegetative shoot growth and thus, return bloom and yield.
The setting ON crop of fruit has a localized and whole tree effect on return bloom
On both ON- and OFF-crop avocado trees, some Year 1 spring shoots set fruit (+fruit) and others fail
to set fruit (-fruit). Clearly ON-crop trees have more shoots that set fruit (+fruit) than OFF-crop trees.
In Year 1, four shoots (one in each quadrant of the tree) that set fruit (+fruit) and four shoots that did
not set fruit (-fruit) on ON- and OFF-crop ‘Hass’ avocado trees were tagged. At bloom the following
year, the number of floral shoots produced by each tagged shoot was counted. Year 1 shoots (+fruit)
on ON-crop trees produced the least floral shoots (Table 2). Year 1 spring shoots (-fruit) on ON trees
and Year 1 shoots (+fruit) on OFF trees produced an intermediate number of floral shoots. Year 1
shoots (-fruit) on OFF trees produced the most floral shoots at return bloom. Note in Table 2 that for
shoots (-fruit) on OFF-crop trees, it is the summer (and fall) shoots that contributed the majority of
floral shoots to return bloom.
The number of shoots that do not set fruit during the ON-crop year is important
Table 2 data clearly illustrate the interaction between the localized effect of fruit present on a shoot
and the whole tree effect related to crop load. These interactions are strongest and least variable for
buds on Year 1 spring shoots (-fruit) on OFF trees and Year 1 spring shoots (+fruit) on ON-crop trees.
The data reveal that the full negative effect of fruit on return bloom in alternate bearing is expressed
on spring shoots that set fruit (+fruit) on ON-crop trees. Note that for ON- or OFF-crop trees, the
greater intensity of the return bloom, and hence the yield obtained in Year 2 will be proportional to the
number of shoots that do not set fruit (-fruit) in Year 1. This is an important concept for managing
alternate bearing. Both Growers A and B have 400 fruit per tree. Grower A’s crop will be followed by
an OFF bloom and OFF crop but Grower B’s crop will be followed by another crop of 400 or more
fruit per tree. Why? Because Grower A has very few shoots without fruit (-fruit) that can produce
vegetative shoots in summer and floral shoots the next spring, whereas Grower B has more than 400
shoots without fruit (-fruit) that will produce summer (and fall) vegetative shoots and a strong bloom
with determinate floral shoots in addition to indeterminate floral shoots the following spring.
The effect of the ON crop on return bloom is cumulative
A fruit removal experiment was conducted to determine when the ON crop is exerting its effect on the
return bloom. This basic information is necessary for timing treatments to mitigate alternate bearing.
The results demonstrated that the effect of the ON crop on return bloom is cumulative. In this
experiment, all fruit were removed from sets of ON-crop trees in progressively later months of the
year. The intensity of the return bloom on these trees was compared to OFF- and ON-crop control
trees. For this experiment, only the setting ON crop of fruit was present on the trees, i.e., the previous
year’s mature crop had been harvested before June. Removing all Year 1 young developing fruit from
ON-crop trees in June, July or August significantly increased the number of summer and fall
vegetative shoots that developed and the intensity of the Year 2 bloom and yield to values greater than
or equal to the return bloom and yield of Year 1 OFF-crop control trees, which were now ON-crop
trees in Year 2. Removing fruit progressively later (from September through December), reduced the
number of summer vegetative shoots that developed, followed by a reduction in the number of fall
vegetative shoots that developed. This reduced the respective contributions of summer and fall shoots
to spring bloom compared to OFF-crop control trees or ON-crop trees with all fruit removed in June,
July or August. Removing fruit as late as December increased the number of floral shoots contributed
by the Year 1 spring shoots during bloom in Year 2 compared to ON-crop control trees with no fruit
removed, demonstrating that Year 1 spring shoots had viable floral buds through December. However,
the number of floral shoots that developed on ON-crop trees with the fruit removed in December was
significantly less than OFF-crop control trees and ON-crop trees with their fruit removed earlier in the
year, which had produced summer and/or fall vegetative shoots that contributed to return bloom.
Moreover, in this experiment, there was an incremental decrease in floral shoot number on Year 1
spring shoots for each month the ON crop remained on the trees past December compared to OFF-
crop control trees and ON-crop trees with all fruit removed in December or earlier. Moreover, leaving
the ON crop of mature fruit on the tree through the summer and fall (until late October) of Year 2 had
a negative effect on return bloom and yield equal to that of the young developing fruit of the ON-crop
(Fig. 3). Having two crops of fruit on the tree increases the total number of shoots with fruit, reducing
the number of summer vegetative shoots that develop in Year 2 and thereby reducing the return bloom
of Year 3!
Figure 3. Holding the mature ON crop of fruit on the tree through the summer or fall of Year 2 reduces summer and fall
vegetative shoot growth of Year 2, respectively, which will reduce the Year 3 bloom, resulting in back-to-back OFF crops.

Year 1 mature avocado fruit do not influence flower and fruit abscission in Year 2
The fact that maturing ‘Hass’ avocado fruit remain on the tree past bloom and can be held on the tree
through October in California raised the question of whether the number of mature fruit on the tree
influences the number of reproductive structures (flowers and young fruit) that abscise (drop off)
during the following year. Our results (Garner and Lovatt, 2008) demonstrated that the OFF- or ON-
crop status of the tree did not influence the percent fruit set or average fruit diameter or weight of
individual fruit that dropped at similar stages of development. Furthermore, the number of mature fruit
in the Year 1 crop did not influence the number of flowers or young fruit that dropped in the Year 2
crop. In both years of the research, as immature fruit drop declined, abscission of the preceding year’s
crop (mature fruit) increased, indicating the processes were controlled independently and not driven by
crop load. The results provided clear evidence that the previous year’s crop load does not influence
current year flower or fruit drop of ‘Hass’ avocado. Thus, the seasonal shedding of reproductive
structures is an independent process that, in the absence of climatic triggers, is not a factor
contributing to alternate bearing.
Strategies growers can use now to mitigate alternate bearing
In many of foreign countries, ‘Hass’ avocado growers are able to use plant growth regulators (PGRs)
not available to California growers to manipulate flowering and yield. For California avocado growers
to remain competitive, yield following the heavy ON crop must be increased substantially to even out
successive annual yields and to stabilize grower profitability. California avocado growers prefer a
cost-effective strategy that does not require fruit removal during the ON-crop year to mitigate the
negative effects of the ON crop and increase yield the following year to sustain their farms and the
industry. Growers rightfully consider flower and fruit removal risky and expensive. Thus, we are
currently conducting research supported by the California Avocado Commission to develop PGR
strategies that mitigate the effect of the ON crop to increase return bloom and yield the following year
with the goal of not having to remove fruit during the ON-crop year. Our PGR research is long-term.
In July 2012, we will begin testing the efficacy of foliar- and irrigation-applied PGRs to mitigate
alternate bearing by treating trees setting an ON-crop, with the goal of increasing the floral intensity of
the putative OFF bloom in spring 2013 and the yield of the 2013-2014 crop. The success of the
treatments will be evaluated after the harvest in July 2014.
In the meantime, growers can take actions that will mitigate the effects of the ON crop to improve
annual income. Please note that the more effective strategies are also more risky. Both the benefit and
risk depend on the degree to which a strategy is imposed and how early in the fruiting cycle that action
is taken. For example, reducing floral shoot number is very risky versus early harvest of fully mature
fruit. In deciding when to act, it is important to remember that the results of the fruit removal
experiment provided evidence that the negative effect of the ON crop on return bloom and yield is
cumulative. Thus, while delaying fruit thinning may be safer with regard to the current crop (Year 1),
the negative consequences on the Year 2 bloom and yield are clear. For many strategies, there is no
proven formula for implementation. Growers will have to learn by doing. The information above and
discussion below provide growers with the critical background information and guiding principles to
make wise decisions in managing their alternate bearing acreage.
Reducing bloom going into an ON-crop year. If your trees need to be pruned, selectively pruning
branches without fruit, to the degree possible, during the winter of the OFF-crop year will reduce the
intensity of the putative ON-bloom while preserving the light OFF-crop. By waiting to prune during
the actual ON bloom, it is easier to gauge the degree of floral shoot removal required and to distribute
the remaining floral shoots, and resulting crop, more uniformly over the tree while preserving the OFF
crop. Results of on-going research in New Zealand suggest that three out of every ten floral branches
could be removed in the ON-bloom year to reduce crop load in the ON year and thus, increase the
number of floral shoots the following year (Jonathan Dixon, personal communication).
Strategies for thinning and harvesting the ON crop. Thinning (fruit removal by hand) the ON crop
prior to summer vegetative shoot growth is the most effective time for increasing return bloom and
yield in Year 2 (Fig.4A). Summer fruit thinning is also the optimal time for increasing the size of the
young developing fruit. Thinning the ON crop prior to fall shoot growth will increase return bloom
and yield, but not as much as summer fruit thinning (Fig. 4B). The Year 1 ON crop has to be thinned
by December (typically before legal maturity) of Year 1 or earlier to have a positive effect on Year 2
bloom and yield (Fig. 4C). Removing as much of the mature (Year 1) ON-crop as possible as early as
possible by size picking combined with harvest of the remaining crop before summer vegetative shoot
growth (July†) reduces the probability of having two OFF crops in a row (Fig. 3). Neither strategy
influences the Year 2 bloom or yield.

Figure 4. A. Thinning (fruit removal by hand) the ON crop in early summer has the greatest potential to increase the number
of summer and fall vegetative shoots that develop and thus, the greatest potential to significantly increase floral intensity and
yield above that of the expected Year 2 OFF bloom and OFF crop. Summer fruit thinning is also the optimal time for
increasing the size of the young developing fruit. B. Thinning the ON crop in fall will reduce the number of summer
vegetative shoots that develop and their contribution to return bloom and yield, but will preserve the growth of fall vegetative
shoots and their contribution to increasing Year 2 bloom and yield above the anticipated OFF bloom. C. Waiting until
December to thin the ON crop will eliminate the contribution that summer and fall vegetative shoots could make to the Year
2 bloom, but will increase the contribution made by the Year 1 spring vegetative shoots so that bloom and yield in Year 2 will
be slightly greater than the putative OFF bloom and OFF crop.

The setting fruit of the ON crop can be removed by hand or by pruning. In either case, the goal is to
reduce the number of fruit in the ON crop uniformly over the tree sufficiently early to promote
summer vegetative shoot growth. Fruit removal by hand is the best and safest approach. Clusters of
fruit can be thinned to increase fruit size, small fruit that are not likely to reach commercial size can be
removed, damaged fruit can be eliminated, and fruit removal can be done so that the remaining fruit
are uniformly distributed over the tree. Research in a commercial ‘Hass’ avocado orchard in
Carpinteria, California, provided evidence that fruit drop of the young, developing fruit is minimal by
the third week of August in both ON- and OFF-crop years (Garner and Lovatt, 2008). If fruit are being
removed by hand, it is possible to wait until this time to thin the crop, which will reduce the risk
associated with high temperatures that can occur during the June drop period (mid-June through the
end of July) and increase return bloom to a greater degree than fall fruit thinning would accomplish. If
the crop is to be thinned by pruning, prune with caution. The goal is a light pruning for crop
management to promote summer vegetative shoot growth. It must be done sufficiently early (spring or
early summer) so the new vegetative shoots have time to mature and develop floral buds. Fall pruning
will not significantly increase return bloom. Pruning off summer shoots from spring shoots that did
not set fruit (shoots without fruit) defeats the purpose. Start small and see how your trees respond.
There are rumors that trees pruned in July or August developed no floral shoots on the regrowth at
spring bloom. Avoid over-pruning, which will interfere with floral bud development. When deciding
how many fruit to remove or how much to prune from ON-crop trees, keep in mind that next year’s
crop will be produced predominantly on Year 1 (current season) summer shoots produced by spring
shoots that did not set fruit (without fruit) or from which fruit have been removed by hand or pruning.
In addition to pruning to reduce crop load, pruning should be done at other times of the year to
increase shoot number (the amount of fruiting wood) and branching (tree complexity), which
contribute to increased floral shoot number and yield the following year.
Fertilizer strategies. Our understanding of best management fertilization practices (BMPs) for an
alternate bearing ‘Hass’ avocado orchard would definitely benefit from additional research. However,
there are practices with known benefits that should be considered. During the OFF bloom, canopy-
applied boron [1.3 lbs/acre B as 6.3 lbs/acre Solubor (20.5% B)] or nitrogen [23 lbs/acre N as 50
lbs/acre low-biuret urea (46% N, 0.25% biuret)] at the cauliflower stage of inflorescence development
can be used to increase fruit set. For ‘Hass’ avocado, a canopy application of B or urea-N at the
cauliflower stage of inflorescence development significantly increased the number of viable ovules,
increased the number of pollen tubes that reached the ovule, and increased total yield and yield of
commercially valuable fruit (packing carton sizes 60 + 48 + 40) (Lovatt, 1999). The application is
made when 50% of the trees in the block have 50% of the bloom at the cauliflower stage, 25% will be
at an earlier stage of inflorescence development and 25% will be approaching or at full bloom (open
flowers). Earlier (bud break) applications were not effective, later (full bloom) applications were
intermediate in effect. Boron is also known to stimulate cell division and increase fruit set and fruit
size of many crops, even seedless fruit, and even when leaf analyses indicate B is adequate. Boron is
preferred over urea-N due to the potential negative effect that can occur when air temperatures reach
90 °F or above after spraying. Late afternoon or early evening sprays are best; morning sprays are fine
as long as the temperature remains moderate throughout the day. The spray should target the
inflorescences. Boron is more effective when applied directly to the flowers; urea is not taken up by
mature leaves of avocado trees grown in California but is taken up by flowers (Nevin et al., 1990). Do
not apply both boron and urea together in the same spray! The effect of applying boron and urea
separately in the same season is not known.
Growers should fertilizer their trees to meet the total demands of the tree throughout the development
of the fruit, including maturing fruit, floral shoots, setting fruit, young fruit during exponential growth,
and summer vegetative shoots, which contribute the greatest number of floral shoots to bloom. Not
only is matching the amount of fertilizer applied to tree nutrient demand cost-effective, it also protects
the environment, especially our water resources. Early research (Lovatt and Witney, 2001) provided
evidence that ‘Hass’ avocado trees in California required 90 lbs N per acre to produce 11,000 lbs of
fruit per acre with a 10% increase in vegetative shoot and root growth and sufficient N to
accommodate volatilization and leaching. The calculation did not take into account the N requirement
of the Year 2 bloom or young developing fruit. Thus, if used as a guide for fertilization, it under
estimates N demand. With recent ON crops exceeding 11,000 lbs per acre, N might be too low during
the ON year in many orchards to support adequate summer vegetative shoot growth for a good return
bloom. Consistent with this, applying 50 lbs of N (as ammonium nitrate) per acre to the soil during the
period of full bloom, fruit set, and vegetative shoot growth (~mid-April) reduced alternate bearing and
increased yield and fruit size of ‘Hass’ avocado trees receiving this treatment for the 4 years of the
experiment compared to trees receiving only 25 lbs of N per acre at this time (Lovatt, 2001). Similarly,
25 lbs of N (as ammonium nitrate) applied to the soil in July and August, when exponential fruit
growth, summer vegetative shoot growth and transition of buds from vegetative to floral development
occur, significantly increased total yield and yield of commercially valuable fruit (packing carton sizes
60 + 48 + 40) compared to trees receiving only 25 lbs per acre in August. The N fertilizer rates
reported here are not prescriptive, but illustrative of the yield benefit derived by meeting tree nutrient
demand at these key stages in ‘Hass’ avocado tree phenology compared to not meeting the tree’s N
demand at these times. The amount of fertilizer to be applied is dependent on yield, canopy size, and
tree and soil nutrient reserves based on leaf and soil nutrient analyses. In addition, irrigation water can
contain significant amounts of essential nutrients, which should be factored into a grower’s
fertilization program. Note that the fertilization goal for the ON-crop year is to provide adequate
fertilizer to support the developing crop and support summer vegetative shoot growth.
Recent research (Rosecrance, Faber and Lovatt, 2011) showed that ‘Hass’ avocado fruit number (crop
load) drives N uptake, as well as the uptake of other nutrients, which parallel fruit dry weight
accumulation. The tree takes up 50% of its total N from full bloom to approximately October, takes up
little N during most winters, and then takes up the remaining 50% the following spring from bud break
(~ mid-February) until just before harvest. With funding from the California Department of Food and
Agriculture Fertilizer Research and Education Program, Dr. Richard Rosecrance, Professor of Plant
Science, California Sate University-Chico, is creating a web-based decision support system (DSS) that
will enable growers to calculate the fertilizer needs of their trees based on the yield of the mature and
setting crops in the orchard, tree size, planting density, tree nutrient status and additional factors. In the
meantime, by estimating the yield of the Year 1 crop of mature fruit and yield of the Year 2 crop of
setting fruit and using the rate of 90 lbs of N per 11,000 lbs of fruit per acre as a reference, one can
calculate the approximate total N that will be required to produce each crop, keeping in mind that 50%
should be applied to the Year 1 mature fruit from bud break until just before harvest, whereas the Year
2 setting crop should receive 50% of the total N required beginning at full bloom through October,
with a significant portion (a minimum of 20-25 lbs) of the total N applied in July and again in August
to support exponential fruit growth and summer vegetative shoot growth. Because both the mature and
setting crops are being fertilized at bloom until just before harvest of the mature crop, trees will be
receiving the extra N required during the period of full bloom, fruit set, and vegetative shoot growth
(~April). A similar approach can be used for other nutrient elements in addition to N. Since fruit
number drives the uptake of many essential nutrients, matching fertilization rates and application times
to periods of high demand by the fruit, as well as periods of strong canopy growth (vegetative and
floral shoot growth) and root growth is a sound approach to fertilization based on tree physiology,
fertilizer-use efficiency, cost to benefit, and protection of the environment.
We are all in this together
In this day of social networking, it is hoped that growers will share their experiences implementing
these strategies, and others, and the outcomes obtained. Until quantitative data is available, grower
experience will be invaluable and it might prove to be the best means to gain knowledge that
adequately reflects the differences in tree architecture and yield potential under the various growing
conditions and management practices used in avocado production in California.
† Phenological stages are provided when possible to assist growers in understanding the timing of
developmental events independent of the effects of local climates. Months, typically provided in
parentheses, are approximations of when phenological events occur but unfortunately are not
applicable to all avocado-growing areas of the state of California.
Relation of carbohydrate reserves with the forthcoming crop, flower formation and
photosynthetic rate, in the alternate bearing ‘Salustiana’ sweet orange (Citrus sinensis
L.)
C. Monerri, A. Fortunato-Almeida, R.V. Molina, S.G. Nebauer∗, A. García-Luis, J.L. Guardiola1
Abstract
The aim of this work was to assess the relation between carbohydrate levels and flower and fruit
production, as well as the role of carbohydrates on CO2 fixation activity, by analysis of leaves, twigs
and roots from the alternate bearing ‘Salustiana’ cultivar of sweet orange (Citrus sinensis [L.]
Osbeck). A heavy crop load (on year) did not affect photosynthesis activity when compared to non-
fruiting trees (off year). Fruiting trees accumulated most of the fixed carbon in mature fruits, whilst no
accumulation was observed in roots before harvest. Non-fruiting trees transported part of the fixed
carbon to the roots and mobilize it for growth processes and, at the end of the season (December),
store it as reserves. Reserve carbohydrates accumulation in leaves started by early December for both
tree types, showing the same levels in on and off trees until spring bud sprouting. A heavy flowering
after an off year caused the rapid mobilization of the stored reserves, which were exhausted at full
bloom. We found no evidence on carbon fixation regulation by either fruit demand or carbohydrate
levels in leaves. Carbohydrate reserves played little or no role over fruit set, which actually relied on
current photosynthesis.
Keywords: Carbohydrate reserves, Citrus, Flowering, Fruit growth, Fruit set, Photosynthesis
1. Introduction
Under subtropical climate conditions, Citrus trees accumulate carbohydrate reserves during the winter
rest and mobilize them during the spring flush of growth (Goldschmidt and Koch, 1996). This
behavior is similar to that described in deciduous fruit trees, which accumulate carbohydrate reserves
before leaf fall and utilize them during the dormant season and the spring growth (Schaffer et al.,
1999). In deciduous trees, the root system is the major storage organ for carbohydrates (Loescher et
al., 1990). In evergreen Citrus, the root system may be the major storage organ for carbohydrates, but
high concentrations of carbohydrates can also be found in the leaves (Goldschmidt and Golomb,
1982). Winter respiration and the beginning of both vegetative and, in some deciduous species,
reproductive growth, occur in the absence of photosynthesizing leaves, and must be totally dependent
on reserves (Loescher et al., 1990). On the contrary, photosynthesis can proceed in Citrus during
winter at a rate high enough to affect growth significantly (Syvertsen et al., 1997; Goldschmidt, 1999).
Therefore, carbohydrate reserves may not be as critical for winter and spring growth in Citrus as in
deciduous trees.
The accumulation of reserves is inversely related to crop load (Goldschmidt and Golomb, 1982), and a
depletion of them under heavy crop load has been related to tree collapse (Smith, 1976) and the
triggering of an alternate bearing habit (Monselise and Goldschmidt, 1982; Guardiola, 1992; Syvertsen
and Lloyd, 1994). Although flower formation can be correlated in some experiments with the
accumulation of carbohydrates (Smith, 1976; Goldschmidt and Golomb, 1982), carbohydrate levels
are not the sole factor regulating flower formation (Goldschmidt, 1999; García-Luis and Guardiola,
2000). During flower formation and fruit set, part of the reserves are translocated to the reproductive
organs (Akao et al., 1981), but the contribution of the reserves to these processes must vary widely as
the rate of their depletion varies among cultivars (Borrás et al., 1984; González-Ferrer et al., 1984),
and differences in reserves depletion rates have also been reported within a cultivar (García-Luis et al.,
1988; Ruiz and Guardiola, 1994; Ruiz et al., 2001). The rate of depletion has been related to flower
number (García-Luis et al., 1988) and there are some studies noting the significance of reserves in
alternate bearing Citrus trees (Goldschmidt and Golomb, 1982) and differences of starch content
between cultivars in relation to crop load (Borrás et al., 1984). However, a lack of relation between
flowering and carbohydrate contents in leaves has been shown in various Citrus species and cultivars
(García-Luis et al., 1988; Lovatt et al., 1988). Although a minimal amount of carbohydrates is required
for bud sprouting and flower initiation, data from several studies suggest that the occurrence of other
controlling factors may obscure the existence of this correlation, indicating that the flowering–starch
relationship remains unproved (Goldschmidt et al., 1985; García-Luis et al., 1988; Martínez-Fuentes et
al., 2010). Furthermore, differences on carbohydrates reserves between early and late maturing
alternate Citrus cultivars could be found.
In the present report, we have studied the seasonal variation of carbohydrate reserves in the leaves, the
twigs (both in the bark and wood) and the roots, in an early maturing alternate cultivar, as well as the
rate of mobilization during the spring flush of growth, in relation to crop load. The objectives of these
studies were: (i) to assess the importance of reserves for the development of the forthcoming crop, and
their role, if any, in the initiation of an alternate bearing cycle; (ii) to analyze the regulatory role of
carbohydrates in flowering and how the changes in carbohydrate reserves through the year reflect the
variations in supply and demand on the diverse organs.
It has been suggested that carbohydrate accumulation may interfere with photosynthesis by feedback
inhibition (Goldschmidt and Koch, 1996), a suggestion supported by the inhibition seen in girdled
branches in the absence of a sink outlet (Schaffer et al., 1986). However, this effect is controversial
under regular cropping conditions (Goldschmidt and Koch, 1996).
Another objective of the present work was to study the effect of crop load and leaf carbohydrate levels
on CO2 fixation rate. The measurements were performed in the ‘Salustiana’ cultivar of sweet orange.
This cultivar has a strong alternate bearing behavior, which allowed us to study the above-mentioned
parameters both in trees without fruit and in trees with a heavy crop load.
2. Materials and methods
2.1. Plant material
The experiments were performed on 35-year-old trees of the ‘Salustiana’ cultivar of sweet orange
(Citrus sinensis [L.] Osbeck) grafted onto Troyer citrange (C. sinensis [L.] Osbeck×Poncirus trifoliata
Raf.) rootstock, grown at Museros, Valencia, Spain. The trees of this parthenocarpic (seedless) cultivar
have an intense alternate bearing habit. In the year 1 of our experiments (2001), some of the trees in
the on year formed a huge number of flowers, resulting in a heavy crop load, which was followed by a
year 2 (2002) with very few or none flowers formed. Some other trees (off trees) presented the same
fruiting habit but shifted by one year, forming almost no flowers and having a low yield at year 1, but
forming many flowers the following one. Under our experimental conditions, bud sprouting of the
spring flush of growth occurred from mid February on fruitless trees to early-mid March on fruiting
trees. Vegetative shoots (bearing only leaves), leafy inflorescences (bearing flowers and leaves) and
leafless inflorescences (bearing only flowers) were formed. Flower opening started by April 1st. Two
waves of fruit abscission were produced from April to June. The mature fruits were harvested by early
February the following year.
2.2. Experimental design
Shortly before flower opening, on and off trees were selected. For the sake of clarity, the trees are
named in the text as on or off according to the number of flowers they formed in the first year of the
experiment. In these two groups of trees we determined the photosynthetic CO2 fixation rate and
carbohydrate accumulation in the leaves, the bark and the wood of the vegetative sprouts formed
during the spring flush of growth, and in the roots. These changes were related to the use of
carbohydrates in fructification, which was determined for the on trees during the first year of the
experiment.
2.3. Spring flush of growth and dry matter use in fructification
A number of vegetative shoots formed during the preceding year and at least 1000 axillary buds were
selected before flower opening in each of six trees of each group in year 1. A random design
comparing on versus off trees was used. The numbers of vegetative shoots, leafy inflorescences and
leafless inflorescences, as well as their number of leaves and flowers were measured during the spring
flush of growth. The number of abscised flowers and fruits were counted at 10–12 days intervals and
abscised organs were collected with a plastic net placed under the tree canopy. By adding to this value
the number of mature fruits at harvest, we obtained the total number of flowers formed. From these
figures, the number of mature fruit present on the tree at any time from flower initiation until harvest,
and the number of shoots and leaves formed during the spring flush of growth, were calculated. Values
are averages of six trees.
The use of dry matter in fruiting was calculated from the number of mature fruits at the end of each of
the measurement periods and the amount of dry matter accumulated by the mature fruits during that
period. The increase in weight of the fruits was calculated as the difference in dry weight of the fruits
at the end and the beginning of each time interval. Taking into account the dramatic upsurge in the
total respiration rate of the flower towards and during anthesis (about 40% of the total dry weight of
the flower according to Bustan and Goldschmidt, 1998), in the first two samplings the calculated
values were corrected for these respiratory losses. The weights were calculated by measuring the
diameter of 200 ovaries/fruits selected at random. The relationship between fruit diameter and fruit
weight was determined in random samplings performed each date on non-measured trees of the same
characteristics, in order to avoid any alteration in fruit number.
2.4. Carbohydrate contents in leaves, twigs and roots
Carbohydrate contents in leaves, twigs and roots both from on and off trees were determined in 10–15
different dates along the year. For each date and treatment (on vs. off) the carbohydrate content
determination was performed from three samples. Two different trees were used to obtain each
sample.
Soluble sugars and starch levels in vegetative shoots leaves were determined from June of year 1, to
the end of June of year 2. From June to September of year 1 three independent samples of 50 leaves
each were collected from the mid portion of vegetative sprouts, both from on and off trees. From
October until late June on year 2, the apical portions (having the five most apical leaves) of ten twigs
formed in year 1 (from two different trees per sample) were sampled. The twigs were separated into
their component parts (leaves, bark and wood) and analyzed separately. New formed twigs during year
2 were not included in the samples. The sampling procedure, sample handling and carbohydrate
determinations were performed as described by Ruiz and Guardiola (1994). Soluble sugars and starch
were extracted with hot 80% ethanol and 35% perchloric acid, respectively, and measured by using the
anthrone reagent.
Roots were sampled at a soil depth of 10–25 cm from positions close to a drip emitter (five positions
per tree). From each tree ten fibrous roots, with a thickness of 1.5–3 mm, were sampled. After careful
washing to remove the soil, the fibrous roots were separated from the recently formed feeder roots.
These two root samples were analyzed separately. The analytical procedure for carbohydrate analysis
was the same described for the leaves.
2.5. Determination of leaf gas exchange
The rate of net CO2 fixation (AN) was measured in both on and off trees during a period (from March
to August) that includes developmental stages with the highest rate of dry matter utilization in the on
tress, that is to say, by March–April when flowering is going on, and June–August when the fruit
growth rate is the highest. AN was measured from June to August of year 1, and from March to May
during year 2, in leaves of vegetative shoots of the spring flush of growth formed during year 1. The
measurements were performed on attached leaves exposed to the sunlight in ambient conditions, with
an LCi Portable Photosynthesis System (ADC, Herst, UK). During the measurements, the leaves were
held perpendicularly to sunlight. Measurements were taken on 7–8 different days each month from ten
on trees and ten off trees. The CO2 assimilation rate for a tree in a day was the mean value of three
consecutive measurements on the same leaf. The CO2 assimilation rate for a tree type in a day was the
mean value from 10 trees. The measurements were performed between 10 AM and 12 PM, to avoid
the midday drop in photosynthesis. The measurements from leaves of the two tree types were
intercalated, to ensure the similarity of the environmental conditions.
2.6. Tree girdling
To determine the influence of the transport to the root system on carbohydrate distribution within the
plant, on and off trees were girdled at the base of the scaffold branches by cutting the bark with a
single edge knife. This procedure removed a 2 mm thick ring of bark. Carbohydrate accumulation in
the leaves and in the bark from vegetative sprouts was determined 6–9 weeks after girdling. No
girdled trees served as controls. Three branches were used in each treatment. The same procedure was
repeated four times on different trees, resulting in a total of 2 tree types (on–off)×4 dates of
measurement (June, September, October and November)×2 treatments (girdling–no girdling)×three
branches of different trees per treatment.
2.7. Statistical analysis
Results were subjected to an analysis of variance (Statgraphics for Windows version 5.1) and means
were separated using LSD multiple range test. To ensure the uniformity of the variance, the
transformations log (X + 1) or arcsin of the square root were used as appropriate.
3. Results
3.1. Characteristics of the spring flush of growth
The characteristics of the spring flush of growth are presented in Table 1. During the first year, the
percentage of bud sprouting was more than double (P≤0.01) in the on trees (56.3±2.7%) than in the off
trees (22.4±0.3%). This resulted in a much higher number of total (leafy + leafless) inflorescences
initiated (20.6 vs. 0.8 thousands per tree, respectively; P≤0.01) and of flowers formed (75 vs. 2
thousands per tree; P≤0.01). On the contrary, the number of vegetative sprouts initiated was 5-fold
higher (8.2 vs. 1.6 thousands per tree) on the off trees than in the on trees. The number of leaves
initiated in spring was similar in both on and off trees. In the off trees, most of the 60 thousand new
leaves were on vegetative sprouts; in the on trees, about 80% of the 49 thousand leaves formed were
located in leafy inflorescences. In addition, individual leaf area was higher (P≤0.05) in the off trees
(30±1cm2 per leaf) than in the on trees. In the on trees, the individual leaf area was similar in the
vegetative sprouts (26±1cm2) and in the inflorescences (23±1cm2).
Tabla 01
Flowering and yield parameters of the trees used in this study.

Parameter and year On trees Off trees

First year
Bud sprouting (%) 56.3 a 22.4 b
Shoots initiated in spring (thousands tres-1)
Vegetative 1.6 b 8.2 b
Leafy inflorescences 6.4 a b
Leafless inflorescences 14.2 a 0.3 b
Number of flowers (thousands tree-1) 75 a 2.0 b
Number of leaves (thousands tree-1) 49 a 60 a
Fruit set (units tree-1) 1,700 a 124 b

Second year
Number of flowers (thousands tree-1) 3.5 b 63 a
Values within a line with different letters are statistically different (P < 0.05).
The number of old leaves (up to 1-year-old at flowering time), could also be calculated from our
flower counts. Meanwhile the on trees formed 75 thousand flowers and the flower count (in thousands
of flowers per 100 nodes) was 0.17±0.03 (average value of six trees), the number of old leaf nodes
present was close to 44 thousand per tree ([75/0.17]×100). A similar calculation for the off trees
yielded 33 thousand old leaves (calculated from a flower count of 0.006 flowers per 100 nodes and a
total flower number of 2 thousand per tree).
During the second year of the experiment, the number of flowers formed was high in the off trees
(which were then in an on year), and low in the on trees (which were in an off year; Table 1). These
figures were obtained from the counts of abscised flowers and fruits, and of mature fruit. The
characteristics of the spring flush of growth (% bud sprouting, number of vegetative and reproductive
shoots initiated in spring and number of leaves) were not measured.
3.2. Use of dry matter in fruiting
In the on trees, fruit abscission occurred from anthesis until the end of June with two distinct peak
values, shortly after anthesis (during April) and by the end of May. The fruit dropped during the first
wave of abscission showed little growth (data not shown). During the second wave of abscission, the
rate of accumulation of dry matter in the fruit increased gradually, reaching a value close to the
maximum by mid June, shortly before abscission ceased (Fig. 1).

Figure 1. Number of surviving fruit (closed dots), and rate of dry matter (DM) utilization in fructification by the on trees
(bars). The least significant differences (P < 0.05) are indicated (LSD) in order to compare the values.

Tabla 2
Dry matter used in fructification (kg tree−1).
Developmental stage On trees (75,000 Flowers) Off trees (2000 Flowers)
Flower formation 14.6 a 0.4 b
From flowering to the end of drop 15.0 a 0.8 b
From the end of drop to harvest 26.1 a 3.1 b
Total 55.7 a 4.3 b

Dry weight in mature fruit at harvest 35.4 a 3.2 b

Values within a line with different letters are statistically different (P < 0.05).
The rate of utilization of dry matter in flower formation and fruit growth was high from bud sprouting
to anthesis, reflecting the cost of flower formation (Fig. 1). Then it fell to a low value, around a fifth of
the previous one, coinciding with the first peak of abscission, to increase gradually and reach the
highest value shortly before the end of abscission, as the rate of accumulation of dry matter in the
mature fruit increased (Fig. 1). After the period of abscission, dry matter use in fruit formation
matched the individual fruit growth, decreasing gradually from September onwards.
During the first year of the experiment, the rate of dry matter utilization in the off trees was in the
range of 3–5% of the values presented for the on trees during flower formation and fruit abscission
and of 11% after fruit abscission (Table 2). The amount of dry matter recovered in the fruit at harvest
ranged between 64% (on trees) and 74% (off trees) of the total dry matter use. The rest was lost
through flower and fruit abscission, and through respiration.
During the second year, the off trees entered an on year and formed a huge number of flowers (Table
1). This year, the rate of dry matter use from flower initiation to anthesis was 256 g tree−1 day−1,
rising at the end of June to 415 g tree−1 day−1. These values are similar to the ones reported above for
the on trees during the first year.
3.3. Carbohydrate content in leaves, bark and wood
The carbohydrate content in the leaves from the vegetative sprouts formed during the spring flush of
growth of the first year is presented in Fig. 2. During the first year, the total carbohydrate content in
the leaves during the final stages of fruit abscission (June) was 2.5-fold higher in the off than in the on
trees. Afterwards, there was a gradual loss of carbohydrates from the leaves of the off trees, and from
September to December the carbohydrate content was lowest, and similar in both tree classes.
Carbohydrate accumulation in the leaves started by early December and proceeded at a similar rate in
both tree classes until the start of bud sprouting in the off trees, by the end of February. In these trees,
which entered an on year, the level of carbohydrates fell to a low value at anthesis (early April),
recovered partially at the beginning of fruit abscission to fall again to the lowest value, close to 10%
(on dry matter basis) by the end of abscission (end of June; Fig. 2). In the on trees, which entered an
off year, carbohydrate content in the leaves was maximal during the period April–May, decreasing by
the end of June (Fig. 2).
The changes in starch content accounted for most of the changes in total carbohydrates in the leaves
(Fig. 2), and the two parameters were closely related (r2 = 0.94; n = 30, P≤0.001). The changes in
soluble sugar content were much smaller and in most of the samples no differences were found
between the two tree classes. At flowering, the sugar concentration in the leaves was higher in the
trees entering an off year than in those entering an on year (Fig. 2).
The pattern of the changes in carbohydrates in the bark and in the wood was similar to that described
for the leaves (Fig. 3). Accumulation in the bark started after November, and no significant differences
were found between the on and the off trees until sprouting. At this time, there was a drastic reduction
in carbohydrates, both in the bark and in the wood, in those trees entering an on year. As for the
leaves, the changes in starch contents accounted for most of these changes (evidence not presented).
The r2 value between these two parameters was 0.94 (bark) and 0.80 (wood).

Figure 2. Changes in total metabolizable carbohydrates, starch and sugars in the leaves of vegetative sprouts formed during
the spring flush in on () and off () trees. The dotted vertical lines indicate the time of harvest. The beginning of bud sprouting
during the second year is indicated by arrows in the axis (S). It occurred earlier in the off trees, which entered an on year,
than in the on trees, which entered an off year. Results expressed as mgg−1 of dry weight (DW). Values are averages of three
independent samples±SE. The least significant differences (P < 0.05) are indicated (LSD) in order to compare the values.

3.4. Effect of girdling on leaf carbohydrates

Girdling the scaffold branches caused the accumulation of carbohydrates in the leaves of the off trees.
This accumulation was much greater during June, October and November, above 30% of the content
in the leaves of the ungirdled controls, than in September, when it was only 14% of the value of the
controls (Fig. 4).
In the on trees, girdling during September and October did not affect the carbohydrate concentration in
the leaves (Fig. 4). When performed in June or in November, girdling increased carbohydrate contents
by 10% above the controls (Fig. 4).
Both in the on and the off trees, the effect of girdling (performed in October or in November) on
carbohydrate contents in the bark of the twigs was similar to that described for the leaves.

Figure 3. Total reserve carbohydrates content in the bark (A) and wood (B) of vegetative twigs formed during the spring
flush in year 1, in on () and off () trees. On year 2, the off trees formed 63 thousands of flowers whilst the on tree formed 3.5
thousands of flowers. Results expressed as mgg−1 of dry weight (DW). Values are averages of three independent
samples±SE. The dotted line indicates the time of harvest; the arrows bud sprouting. The least significant differences (P <
0.05) are indicated (LSD) in order to compare the values.
3.5. Carbohydrate content in the roots
From October to early December carbohydrate content in the roots was the same in the off and in the
on trees (Fig. 5). At this date, a gradual accumulation of carbohydrates started in the roots of the off
trees, both in the fibrous and, in a lesser amount, in the feeder roots (Fig. 5). This accumulation ceased
at the time of bud sprouting, since a significant part of the accumulated reserves were utilized in just
20 days (Fig. 5). At this time there was a significant increase in the proportion of feeder roots (Fig. 6).

Figure 4. The effect of girdling on the accumulation of carbohydrates in the leaves in on (open bars) and off (grey bars)
trees. Results expressed as a percentage of the carbohydrate contents in the leaves of non girdled (control) trees±SE (n = 3).
The asterisks indicate statistically significant differences to the controls at P≤0.05 (*) and P≤0.01 (**).

In the on trees the accumulation of carbohydrates started after harvest and ended by bud sprouting
(Fig. 5). The maximum concentration of carbohydrates in the roots of the on trees was much smaller
than in the off trees. After sprouting the concentration of carbohydrates in the roots was identical in
the two tree classes.
Most of the changes in carbohydrates content reflected the accumulation and the mobilization of
starch, whose concentration was closely related to that of total metabolizable carbohydrates (r2 = 0.98
and 0.90 for the feeder and the fibrous roots, respectively; n = 20; P≤0.001).
Figure 5. Fig. 5. Total reserve carbohydrates content in the feeder (A) and the fibrous (B) roots in on () and off () trees. On
year 2, the off tree formed 63 thousand flowers whilst the on tree formed 3.5 thousand flowers. Results expressed as mgg−1
of dry weight (DW). Values are averages of three independent samples. The dotted line indicates the time of harvest; the
arrows bud sprouting. Values are averages of three independent samples±SE. The least significant differences (P < 0.05) are
indicated (LSD) in order to compare the values.

3.6. Effect of fruit load on photosynthesis

During the periods with the highest rate of dry matter utilization in the high crop trees, that is to say,
by March–April when flowering is going on, and June–August when the fruit growth rate is the
highest (Fig. 1), there were no significant differences in photosynthetic rates between the trees with a
low and a high crop (Table 3). There were not differences between months either. The main factor
affecting changes on this parameter is the variability among days in a month (Table 3). Leaf
photosynthesis was largely determined by the environmental conditions. It is important to point out
that differences in carbohydrates between tree groups are highest during these studied periods (Fig. 3).
Figure 6. Feeder roots weight, expressed as a percentage of the total weight of the sample, in the root samples from on ()
and off () trees. Values are averages of three independent samples±SE. The least significant differences (P < 0.05) are
indicated (LSD) in order to compare the values.

Table 3.
Effect of crop load on the photosynthetic rates (AN) in the leaves of vegetative sprouts of ‘Salustiana’
sweet orange during the months of highest dry matter utilization. The values are average of 6–7
different days per month (10 replicates per day) performed from March to August.

March April May June July August

On trees 7.19 6.85 8.12 8.02 8.19 8.18
Off trees 7.62 7.37 8.31 7.51 6.13 6.74

ANOVA
Source of variation SS DF MS F
Month 18.43 5 3.69 0.61 ns
Crop load 3.48 1 3.48 0.57 ns
Month x crop load 16.27 5 3.25 0.54 ns
Error 473.35 78 6.07
ns: not significant effect.
4. Discussion
It is well known that the fruit is a major and priority sink in Citrus, and that a heavy fruit load reduces
both the diversion of carbohydrates towards the root system (García-Luis et al., 1995) and the
accumulation of carbohydrate reserves (Goldschmidt and Koch, 1996). The amount of the mid-winter
carbohydrate reserves in an off year in an adult tree of the alternate bearer Wilking mandarin has been
estimated at ca. 24 kg (Goldschmidt and Golomb, 1982). These reserves are mobilized during the next
on year (Syvertsen and Lloyd, 1994; Goldschmidt and Koch, 1996), and it has been speculated they
could satisfy a considerable portion of the dry matter requirements of the following crop (Goldschmidt
and Golomb, 1982). However, the contribution of the reserves to the fruit set varies between and
within cultivars (Borrás et al., 1984; González- Ferrer et al., 1984; García-Luis et al., 1988; Ruiz and
Guardiola, 1994; Ruiz et al., 2001) and carbohydrate levels are not the sole factor regulating flower
formation (Goldschmidt, 1999; García-Luis and Guardiola, 2000; Martínez-Fuentes et al., 2010). The
results of the present work, in an early maturing alternate Citrus cultivar, cast doubt on the role that
carbohydrate reserves play in fruit set and no evidence could be found for a regulatory role of
carbohydrates in flower induction and initiation.
4.1. Use of dry matter in fructification
In our experiments the concentration of reserve carbohydrates in the off trees, which entered an on
year, at the onset of bud sprouting was of the same order of magnitude as reported for other alternate
Citrus cultivars at the end of an off year (Goldschmidt and Koch, 1996). These reserves were used-up
during bud sprouting, and at the time of full bloom (early April) the carbohydrate concentration in the
leaves and in the twigs had fallen to a minimum value (Figs. 2 and 3). During the period of fruit
abscission (May and June), in which competition for carbohydrates is considered to be a limiting
factor for fruit retention (Goldschmidt, 1999), fruit nutrition was supported by current photosynthesis
and the carbohydrates stored after anthesis (during April; Figs. 2 and 3). The relatively high
carbohydrate contents in the leaves and in the bark at full bloom, around 10% on a dry matter basis,
may represent a non-utilizable fraction (Goldschmidt and Koch, 1996; Ruiz et al., 2001).
As reported for other woody species (Loescher et al., 1990) the root system was the main storage
organ for carbohydrates. These reserves fell dramatically at the onset of bud sprouting, in particular
those from the fibrous roots (Fig. 5), coinciding with the resumption of root growth (Fig. 6). This fall
in carbohydrate reserves was relatively smaller than in the leaves and in the twigs, and the fibrous
roots of the trees entering an on year had, during anthesis, a carbohydrate concentration twice the
amount found in early October. These root reserves could potentially be utilized for fruit set, but the
net change in these reserves during the period of fruit abscission was low and there were no
differences in the carbohydrate content between on and off trees during fruit set (Fig. 5). The girdling
experiments demonstrated that at this time the root system competed with the fruit for the available
carbohydrates (Fig. 4). Although some transport of carbohydrates from the root system to the shoot
was demonstrated in labeling experiments (Kubota and Motoyama, 1972; Goldschmidt and Koch,
1996), we found no evidence for the contribution of the root reserves to fruit set. A similar conclusion
was reached by Loescher et al. (1990) in their review on root reserves in deciduous trees, as these
authors stated that evidence for the roots playing a special role on fruiting behavior is unclear.
Whilst the reserve carbohydrates may have supported the initial stages of vegetative growth and
reproductive development (Akao et al., 1981) they played no significant role in fruit set. Most of the
reserves stored during the off year may have gone in surplus flower formation, whose cost in term of
carbohydrates was about 60% of the total reserves (Table 2). This surplus flower formation had little
or no effect on fruit set (Becerra and Guardiola, 1984), but in some conditions it may impair it
(Becerra and Guardiola, 1984; Guardiola et al., 1984).
4.2. Role of carbohydrate reserves on alternate bearing habit
It is generally accepted that the cycles of alternate bearing are caused by the inhibition of flower
formation after a heavy fruit load (Guardiola, 1992), and flower formation has been related to
carbohydrate levels (Goldschmidt and Golomb, 1982; García-Luis et al., 1988, 1995). This role of
flower inhibition by the fruit was supported by our results (Table 1) which are in agreement with those
of other works in Citrus (Martínez-Fuentes et al., 2010; Verreynne and Lovatt, 2009), but we could not
find evidences for a regulatory role of carbohydrates. In some alternate Citrus cultivars which are late
maturing, the presence of developing fruit prevents the accumulation of carbohydrates in all tree
organs during an on cycle (Syvertsen and Lloyd, 1994). In the early maturing ‘Salustiana’ orange used
in this study, the low fruit strength of the fruit after November (Fig. 1) allowed the accumulation of
reserves during winter in the leaves and the twigs of the on trees (Figs. 2 and 3), and during the period
of flower induction and initiation (from early November until bud sprouting) the carbohydrate levels
in these organs were identical in the off and the on trees. Therefore, we may conclude that
carbohydrate reserves did not play a regulatory role in flower development, otherwise questioned by
Goldschmidt (1999), García-Luis and Guardiola (2000) and Martínez-Fuentes et al. (2010). As most of
the flowers formed in the axillary buds of the vegetative shoots were formed during the spring flush of
growth of the preceding year, the 5-fold higher number of these buds in the off trees than in the on
trees (8.2 vs. 1.6 thousand vegetative shoots per tree, respectively; Table 1) would explain only part of
the 20–35-fold differences in flower formation between the off and the on trees (Table 1). A direct
effect of the fruit on flower initiation in the buds also seems evident.
4.3. Source–sink effects on seasonal carbohydrate reserves
The changes in carbohydrate reserves during the year reflect the variations in supply and demand. The
carbohydrate demand during the initial stages of vegetative growth and reproductive development in
trees entering an on year produced an initial drop in the leaf and twig reserves which was followed by
a transient recovery and a further drop at the end of fruit abscission (Figs. 2 and 3). In trees entering an
off year with a low carbohydrate demand for fructification, carbohydrate accumulation continued until
early April and remained high until the end of May (Figs. 2 and 3). The carbohydrate levels in the
leaves during spring in trees with a low crop were double than the levels in the high crop trees (Fig. 2).
The differences in the pattern of the changes in carbohydrate reserves reported by different authors
(see references in Section 1), which lead sometimes to high starch levels during May in the northern
hemisphere (Borrás et al., 1984), might be the result of differences in flower formation rather than a
varietal characteristic.
After fruit abscission, carbohydrate levels in the leaves declined and stayed low during summer and
autumn. As no effect of the fruit on photosynthesis (Table 3) or on leaf carbohydrates (Fig. 2) was
found during those months, most of the carbon fixed by the off trees must have been transported to the
root system, a conclusion supported by the girdling experiments (Fig. 4). The smaller retention caused
by girdling in the off trees during September may reflect competition from the summer flush of
growth. This higher transport of carbohydrates to the root system in the off trees did not increase the
root reserves (Fig. 5) nor root growth, as assessed by the proportion of feeder roots to the total root
sample (Fig. 6), and it may have been used in the thickening of the major roots. Interestingly,
Goldschmidt and Golomb (1982) reported a 2-fold higher weight of the major roots after an off year
than after an on year in Wilking mandarin.
Despite the low sink strength of the fruit from December onwards (see the low rate of dry matter
utilization from December. Fig. 1), it prevented the accumulation of carbohydrate reserves in the root,
which only occurred after harvest, an effect previously observed in potted trees (García-Luis et al.,
1995). This fruit effect, whose regulation was not investigated, did not affect the accumulation of
carbohydrates in the leaves (Fig. 2). During this period of time, fruit load affected the carbohydrate
reserves in the roots but had no effect on their concentration in the leaves. As pointed out by
Goldschmidt and Golomb (1982) the leaves are not always the most sensitive indicator organs. In this
study we explained how, depending on the time of sampling, the carbohydrate concentration in the
leaves of the trees with a high crop was lower (March–July) or the same (autumn and winter months)
than in the trees with a low crop.
4.4. Photosynthesis regulation
The regulation of photosynthesis (AN) by sink demand has been demonstrated in Citrus using either
potted trees (Syvertsen and Lloyd, 1994; Goldschmidt and Koch, 1996; Iglesias et al., 2002), or
altering the source–sink balance in the tree (Li et al., 2003). A regulatory role of the carbohydrate level
in the leaves has been suggested (Iglesias et al., 2002). However, to what extent sink demand controls
photosynthetic rates in Citrus under regular cropping conditions is not clear, nor is the mechanisms of
regulation (Goldschmidt and Koch, 1996). We addressed this issue by measuring AN under-orchard
conditions during a period (from March to August) including that developmental stages in which the
differences in the sink strength of the developing flowers and fruits between the high and low crop
trees were highest; the flower formation period (March–April), and the main fruit growth stage (June–
August). The highest differences in leaf carbohydrate contents happened during this period. During
these months the environmental conditions were the main determinant of AN. No enhancement of AN
by a high sink strength related to fruiting was found in our experiments. A delay of leaf senescence
caused by fruiting, and an enhancement of autumn photosynthesis, has been demonstrated in some
works with apple (Tartachynk and Blanke, 2004), but is doubtful whether a similar regulation may
occur in Citrus, whose leaves show a high photosynthetic efficiency during 2 years (Kubota and
Motoyama, 1972). Our data supported the view that photosynthesis in non-manipulated Citrus trees
was source rather than sink limited at least during the months measured in this study (from March to
August). As shown previously for deciduous trees like cherry (Roper et al., 1988) and other works
with apple (Lakso et al., 1999), fruit load had no significant effect on photosynthesis as carbohydrates
may be translocated to alternate sinks. Nor could we find a depressing effect of leaf carbohydrates on
AN, as this parameter was similar in trees with low and high crop when differences in carbohydrate
content were the highest, and the 13–15-month-old leaves of the low crop trees (Table 3) had a very
high carbohydrate concentration (Fig. 2). Interestingly, the carbohydrate concentration in these leaves
(25% on a dry matter basis) was higher than the concentration that Iglesias et al. (2002) reported as
inhibitory for manipulated Satsuma mandarin trees (Citrus unshiu) (ca. 18% on dry matter basis).
Diverse studies on different species (Paul and Foyer, 2001) cast doubt on a simple relation between
starch accumulation and feedback regulation. Some evidences suggest that night-time hexose content
derived from starch may provide signals for feedback regulation through modulation of gene
expression (Cheng et al., 1998). In the work of Iglesias et al. (2002) treatments of a different nature
were carried out and it is difficult to compare the effect of absolute content of sugars when other
physiological parameters could be very different. If photosynthesis is regulated by sink activity or by
carbohydrate content is not clear.
These conclusions may not extend to the full year. Syvertsen et al. (2003) found that during winter
(late June in the southern hemisphere, equivalent to late December in the northern hemisphere) de-
fruited ‘Spring’ navel orange trees had a 40% lower AN value than fruiting (and already harvested)
trees. Although defruiting may cause a transient change in AN (Gucci et al., 1991) and the trees used
by Syvertsen et al. (2003) were smaller in size (5 years old) than the ones used in our experiments, the
possibility that tree behavior during the winter months may be different as described in this report
cannot be dismissed.
In conclusion, the changes in carbohydrate reserves through the year reflected the variations in supply
and demand. Carbohydrates were stored in shoots and roots by winter in non-fruiting trees, and
mobilized during the spring flush of growth. Shoot reserves were used-up during bud sprouting and
flowering, and at the time of full bloom carbohydrate concentration had fallen to a minimum value. No
evidence was found about the contribution of the root reserves to fruit set. Thus, from the period of
fruit abscission, fruit nutrition was supported by current photosynthesis. Fruit load causes the
inhibition of flowering after heavy crop load in alternate bearing trees, but a regulatory role of the
carbohydrates could not be observed. And no effect of leaf carbohydrates or fruit load on
photosynthesis could be found. We understand that fixation rates were mainly modulated by
environmental conditions in Salustiana sweet orange.