Performance of Two Very Early-Season Clementines, ‘Clemenrubi’ and ‘Orogros’ Mandarins on Three Rootstocks in Spain: Yield and Quality Study

Abstract

The yield and fruit quality performances of Clemenrubi and Orogros mandarins were evaluated in the Mediterranean climate of Valencia, Spain, from 2009 to 2016. The cumulative yields and tree growth of both varieties’ mandarins over this 8-year production period were lower in Clemenrubi. Fruit size was affected by the rootstock being the control rootstock (CC), which induced the smallest fruit, although yield efficiency and alternate bearing behaviour were not affected. Although C-35 was highlighted for delaying fruit maturity (lower RI and CI indices), no significant differences in TSS were obtained compared to the other rootstocks. CC also presented strong granulation disorder and C-35 the highest tree mortality. In conclusion, all the rootstocks for Clemenrubi and Orogros mandarins gave good fruit quality for fresh fruit markets in the eastern Mediterranean region. Nevertheless, long-term studies are needed to determine the exact effects of multibud galls on tree survival and the granulation problem on the CC fruit.

1. Introduction

Spanish citrus production areas are environmentally suitable for growing multiple varieties of mandarins (Citrus reticulate L. Rutaceae), which provide diverse production options with a very long marketing calendar. In production terms, Spain accounts for nearly 2 million tons of mandarins, whose main destination is exportation (63.65% in relation to Spanish mandarin production), which places Spain the first exporting country worldwide [1]. Despite advances made in extending the varietal calendar, Spanish markets continue to concentrate vast mandarin production (50% of its total production) in November and December. Hence, establishing cultivars or hybrids that avoid overproduction during the harvest period when prices are low is a desirable fact in new orchards from most citrus-growing countries, including Spain.

Among oranges, early cultivars such as Cadenera and Dahong are highlighted for their good fruit quality (high juice percentage and high titratable acidity (TA), respectively) and Salustiana for its good levels of total soluble solids (TSS) and TA [2]. Among later cultivars, Valencia Delta Seedless is the most interesting cultivar for having the highest recorded yields and TA levels, while other cultivars such as Barberina and Midknight are highlighted for their high TSS and juice contents, and Valencia Rhode Red for its colour index (CI) [2]. ‘Midknight Valencia’ and ‘Rhode Red Valencia’ are also suggested as good late-season orange cultivars in Turkey [3,4]. Among mandarins, planting ‘Okitsu’ and ‘Clausellina’ satsumas (respectively of Japanese and Spanish origin) was an apparently good option in the 1990s for their early season ripening, but their organoleptic quality was very low (poor TSS and TA contents) and tended to present puffiness upon ripening [1,5]. So, their current marketability is being displaced by the higher quality of early mandarin clementines. This group includes many varieties from trees of normal size and with seedless fruit (but can pollinate and be pollinated with compatible varieties), along with good fruit quality which, however, decreases quickly upon ripening and also tends to display puffiness. Clemenrubi and Orogros are two very early-season mandarins with an attractive orange-reddish colour that derive from mutations of clementine ‘Oronules’. Both varieties advance fruit ripening between 10 and 15 days compared to Oronules (http://www.ivia.es/variedades/, accessed on 28 November 2021), which makes their interest from an economic point of view quite clear. However, problems related with yield in these varieties are frequent. Clemenrubi tree is less vigorous than other early varieties [1]. In both cases the appearance of multiple-bud galls on trunks is very frequent, which constrains plant vascular system and generates problems of incompatibility between scion and rootstock. This affection reduces not only the tree size but also the plantation life, because it shortens productive life and generates even serious tree loss problems earlier than in other varieties, especially when they should have reached their optimum productive age [1,6,7]. Although prices are most appealing at the start of the season, serious drawbacks with yields, shorter plantation life and production and quality irregularities make the costs of crops in these varieties very high, and also in early-season ones in general. This problem needs an immediate solution.

Today’s increased search and demand for early and late cultivars worldwide is possibly due to changes in consumer habits and preferences and, consequently, in market requirements. Fruit quality attributes, such as colour, taste (sugar and acid), flavour, and quality-related enzyme effects are extremely important commercially. Moreover, due to increasing consumer interest in health-promoting food properties, aspects related to nutritional value, SST and TA contents, ascorbic acid, and antioxidant proprieties contribute to define fruit quality. It has been indicated that different quality attributes may significantly vary among citrus species, among varieties within the same species, and even within the same variety grown under diverse climatic conditions [8,9].

Nevertheless, fruit quality can also be devalued by the granulation physiological disorder, which affects some citrus varieties, especially mid-season and late oranges, but also some mandarins, particularly when left on trees after ripening and which become worse during post-harvest storage. Fruit vesicles partially dry, normally those next to the peduncle, because cells can lose up to one third of juice content and gel part of their soluble solids, and they acquire a crystalline, hardened and granular appearance, pale colour, and insipid taste, which may make consumers think that fruit is frozen [10]. The origin of this alteration is related to increase fruit respiration, and suggests that produced energy is used in the thickening and modification of cell walls, and these changes increase with fruit ageing. Fruit size, position of fruit on trees, delayed harvesting, sandy soils with little water retention capacity and high temperatures after fruit colour, or temperatures around 0 °C in winter, late blooms, etc., are some factors that influence this physiological disorder [11]. Therefore, it is difficult to fight this alteration, and recommendations are related to the advance of the harvests of those plots or varieties that present a habitual tendency to granulation or the application of some phytoregulators such as gibberellic acid. Once again, plant material seems decisive in the presence and evolution of this anomaly as it is directly related to rootstock vigour [11]. Vigorous stocks induce a bigger quantity of affected fruit and these, in turn, are affected by greater intensity.

It is well-known that rootstocks have played a key role in citrus industry’s development in the world and its utilisation has gained value because it can affect tree behaviour, which are important limiting and restricting factors of citrus production (soil, climate, pests, etc.) and condition market demands in terms of productivity, short juvenility period, and high-fruit quality [12]. Rootstock influences the development of the ripening process, skin colour and juice, SST and TA contents, and other juice characteristics [13]. The selection of an appropriate scion/graft combination should prioritise those rootstocks that provide better fruit quality to meet citrus industry demands. The appearance of Citrus Tristeza virus obliged Spanish citriculture to replace the Sour orange (Citrus aurantium L.) rootstock with other plant materials that were resistant to this disease. This is the case of Carrizo citrange (CC, P. trifoliata (L.) Raf. × C. sinensis (L.) Osb.) and it is the main rootstock currently used in Spain in more than 80% orchards. This trifoliate rootstock is resistant to Phytophtora and psoriasis, confers the variety good yields and quality, but it is susceptible to iron chlorosis, one of the main constraints of calcareous Spanish soils. C-35 (C. sinensis × P. trifoliata) is resistant to nematodes, vigorous and induces good yields. It is more sensitive to limestone than CC but is less sensitive to salinity and well tolerates cold. So, it is suitable when these stresses appear. Finally, Forner-Alcaide 5 (FA-5), a hybrid of Cleopatra mandarin (C. reshni) and P. trifoliata, has gained followers in the last decade (7.2% in Spanish orchards during 2020) for its good behaviour in calcareous soils and its tolerance to salinity, in addition to Phytophtora and nematodes (https://agroambient.gva.es, accessed on 28 February 2022).

The aim of this work was to determine some parameters related to tree growth parameters and multibud galls incidence, as well as fruit quality and the granulation disorder, in two very early-season mandarin varieties (Clemenrubi and Orogros) grafted onto three different rootstocks. These results would help breeders to make decisions about selecting new citrus materials in their orchards, which are more suitable for market demands considering external fruit quality, organoleptic properties, and tree development.

2. Materials and Methods

2.1. Plant Material and Field Trial Design

The citrus rootstocks used in the experiment were CC (control), C-35 and Forner and Alcaide 5 (FA-5). Seeds came from mother seed trees held in the germplasm collection at IVIA (Instituto Valenciano de Investigaciones Agrarias). Scion response was studied on two early-season clementine mandarin varieties (‘Clemenrubi’ and ‘Orogros’). Buds were obtained from the germplasm collection at the IVIA. Rootstock seedlings were budded when they were 12 months old.

All the nursery processes were performed in an aphid-proof greenhouse with a cooling system, and temperature ranged between 18 °C and 27 °C with relative humidity of about 80%. In March 2004, nursery trees were planted in a randomised complete-block design with five blocks (four trees each) for each scion/rootstock combination. The total number of trees was 20 for each scion/rootstock combination. Tree spacing was 6 m × 4 m. The plot was surrounded by border rows on all four sides. Trees were grafted with the scion 1 year later (July 2005).

The field was located at COPAL—Cooperativa Agrícola in the town of Algemesí (about 15 km from the Mediterranean Sea, Valencia, Spain). Soil texture within the first 50 cm depth was classified as clay loam soil, with pH 8.5, CaCO₃ 44.4%, active calcium carbonate 17.1% and electric conductivity in the saturation extract at 25 °C of 5.79 mS cm^–1.

The standard cultural practices for this variety were used with drip irrigation and chemical weed control. The water pH was 7.8, electrical conductivity ranged from 2.0 to 3.5 mS cm⁻¹ mS cm⁻¹, and with 400 to 500 mg kg⁻¹ of B. Fertilisation was applied from year 2 and increased annually. Since year 6, the amounts applied were: ammonium nitrate (33%) 2 kg tree⁻¹, mono-ammonium phosphate 0.5 kg tree⁻¹, KNO₃ 0.7 kg tree⁻¹, and iron chelate 20 gr tree⁻¹. After 3 years of cultivation, trees were hand-pruned annually after harvest.

2.2. Growth Parameters

Seven years after grafting, tree height and canopy diameter were measured for all the trees and scion/stock combination. Canopy volume was calculated using Turrell’s formula [14]. The percentage of tree mortality was calculated at the end of the experiment (2016) and related to the incidence of multibud galls on trunks.

2.3. Yield Parameters

Fruits were sampled when the first commercial harvest took place for 8 consecutive years (from 2009 to 2016). Mandarin samples were collected in late September (between 13 and 27 September). The fruit from each tree were harvested, and the collection of the recently dropped fruit and yield (kg tree⁻¹) was determined. To prevent from possible differences between rootstocks in preharvest drop terms, no 2,4-D treatments were applied and during harvests 6, 7, and 8, the fruit that dropped from 3 weeks before harvest date was added to the final year yield. The weights of these fruit were added to the weights of the harvested fruit. Cumulative yield (kg tree^–1) for the 8-year period was also calculated and yield efficiency (kg m^–3) was estimated as the ratio of cumulative yield to canopy volume.

The alternate bearing index (ABI) was calculated by dividing the difference between two consecutive harvests by the sum of two yields × 100% for the nine harvests. If the index is more than 50%, it means that the tree presents alternate bearing, but regular bearing if this index is less than 50%.

2.4. Fruit Size Determinations

Fruit quality was determined for four harvests: 2010, 2011, 2012, and 2013. At the harvest time, a selection of representative fruit that reached maturity was randomly collected from 12 trees per scion/rootstock combination. Briefly, 20 fruits per tree (5 fruits from each cardinal canopy position) were collected to sum up a total of 240 fruits per scion/rootstock combination. Fruits were mixed together, and four subsamples (25 fruits each) were made. One subsample (n = 25) was selected to measure fruit physico-chemical characteristics. Fruit weight was measured with a digital balance (Sartorius, Model BL-600, 0.01 g accuracy). Fruit diameter (D), fruit height (H), and peel thickness were measured with an electronic digital slide gauge (Mitutoyo, Model CD-15 DC, 0.01 mm accuracy) and the fruit shape index (D/H) was calculated.

2.5. Fruit Internal Quality

Three subsamples (25 fruit each) were used to determine quality characteristics in juice. Fruit was squeezed together (3 replicates for each scion/stock combination, n = 3) with an electric squeezer and the juice percentage (w/w) was calculated. The percentage of fruit with dried vesicles was also determined. TSS were measured with a refractometer (Model N-1, Atago (±0.2 °Brix)) and expressed as percentage at 20 °C. The method to analyse TA was based on neutralisation (NaOH 0.1 N) to pH 8.1. Values were expressed as %. The ripeness index (RI) was calculated as the TSS/TA ratio.

2.6. Peel Colour Index

The colour index was determined in peel according to ref [15]. Briefly, the CIELAB L* (brightness or lightness; 0 = black, 100 = white), a* (−a* = greenness, +a* = redness), and b* (−b* = blueness, +b* = yellowness) colour variables were measured using a chromatometer (Minolta, Model CR-300, Ramsey, NJ, USA) coupled to a Minolta DP-301 data processor. The colour index (CI) was then calculated by applying the following formula CI = 10³ a*/L*b*.

2.7. Statistical Analysis

Statistical analyses were performed by Statgraphic Centurion XVII (Statistical Graphics Corporation 2014) using the rootstocks as the factor of analyses for each variety. A basic descriptive statistical analysis was followed by an analysis on variance test for mean comparisons. Duncan test method was followed to discriminate among means at a 95% confidence interval (p ≤ 0.05).

A principal component analysis (PCA) was run by Statgraphic Centurion XVII software, for the standardized values using pairwise Euclidean distances among means to determine the relations between scion/rootstock combinations as the factor of analyses. The extracted eigenvalues (>1), and the relative and cumulative proportions of total variance explained by the first two principal components (PCs) were calculated. Two two-dimensional (2D) scatter plot (first PC vs. second PC) were prepared based on a distance matrix for the PCs to visualize scion/rootstock relation and trait weight. Additionally, a dendrogram analysis was conducted.

3. Results and Discussion

3.1. Growth Parameters

Among varieties, Orogros trees were the biggest ones as reflected by the visual tree aspect (Figure 1). Some growth parameters (

Table 1. The tree growth parameters (height and canopy diameter, in m; and canopy volume, in m³) of ‘Clemenrubi’ and ‘Orogrós’ Clementine mandarin trees on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) measured 7 years after planting ^a.

Variety	Rootstock	Tree Height a (m)	Canopy Diameter a (m)	Canopy Volume a (m3)
Clemenrubi	CC	1.8 a	2.0 a	4.1 a
	C-35	2.0 a	2.3 a	5.2 a
	FA-5	1.9 a	2.3 a	5.3 a
	Mean	1.9 ns	2.2 ns	4.9 ns
Orogros	CC	2.1 a	2.8 b	8.5 b
	C-35	2.1 a	3.1 a	10.5 a
	FA-5	2.1 a	2.9 ab	9.6 ab
	Mean	2.1 ns	2.9 **	9.5 *

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, the numbers followed by different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and ns (non-significant).

), such as canopy diameter and canopy volume (33.3% and 95.9% higher on average, respectively, than Clemenrubi). The trees on C-35 also presented clear chlorosis symptoms on their leaves (yellowness and defoliated branches) independently of variety (Figure 1).

Among the ‘Clemenrubi’ trees, no statistical differences between rootstocks were found for any parameter related to tree growth (Table 1). In ‘Orogros’, the rootstock did not influence tree height, while the CC trees presented the smallest canopy diameter and canopy volume (4.5% and 9.9% of the mean value, respectively).

The influence of rootstock on plant development is a well-known issue in citrus. According to the bibliography, the three rootstocks herein employed confer oranges, lemons or clementines a medium-high vigorous growth habit [16,17,18]. In a study on ‘Marisol’ Clementine, Bassal (2009) stated that CC and ‘Swingle’ citrumelo induced higher tree height compared to Cleopatra mandarin. In Nova mandarin, CC induced a bigger canopy diameter than other rootstocks [19]. In ‘Navelina’ orange [15], the biggest trees in height and canopy volume terms were obtained when grafted onto Volkamer lemon, CC, C-35 and FA-5, with no statistical differences between them, while the smallest trees were obtained on some hybrids: FA-03027 (Cleopatra mandarin × Troyer), FA-020349 and FA-02034 (Troyer × Cleopatra mandarin) and FA-13 (Cleopatra mandarin × Rubidoux).

So, the clear difference in tree canopy between Clemenrubi and Orogros is related to the influence of variety, but not of rootstock. As general rule, although almost all very early-season varieties recently cultivated in Spain have a common origin (mutations of the Oronules variety), Orogros is the most vigorous one. Differences in growth habit between varieties of the same group or species have also been described independently of rootstock by other authors. In a study that compared tree development in two early-season Satsumas, Clausellina was smaller than Okitsu regardless of rootstock (Sour orange, Carrizo citrange and Troyer citrange; [20]).

3.2. Tree Mortality

The number of trees on all the variety/rootstock combinations had decreased at the end of the experiment. The highest percentages of mortality (Figure 2A) on Clemenrubi trees were found on stocks C-35 and FA-5 (affecting 54% in both cases), which locates these combinations in the upper zone of the plot (Figure 3B). In Orogros, this disorder was relevant on C-35 (50%), while low mortality was observed on CC and FA-5 (15%).

The possible reason for this result is the trend to produce multibud galls in the trunk. Although all the scion/stock combinations presented this disorder, albeit with some differences between rootstocks (Figure 2B). The highest incidence was visually observed in Clemenrubi/CC, Clemenrubi/C-35, Orogros/CC, and Orogros/C-35, especially in the scion/rootstock insertion. Multibud galls appearance is a common disorder in early-season varieties, which is once again likely due to their common origin, and suggests strong genetic regulation, as reported previously [21]. Although the quantity and size of multibud galls are characteristic of variety, certain variability in symptomatology depending on rootstock and age has been observed and scion-rootstock incompatibility may also have an influence [22]. In this sense, cultivation on Carrizo citrange in addition to high planting densities so that early returns could offset losses if incompatibility appeared after 7 years, was recommended by [22]. As gall formation blocks citrus’ vascular system, plants may display poor shoot growth with wilting, yellowing, dramatically reduced yields and death occurring above the gall, especially if galls girdle trunks. Severe damage results in tree death, which is coincident with a necrotic line in the bud union that was the main cause of mortality of Fukumoto/Swingle citrumelo, or can result in dwarfing habit such as Fukumoto/C-35 with trees being at least half the size of the others, but lack the dark ring [22]. The trees on C-35 and Rough lemon produced smaller scions and shorter trees than CC and Valencia/Carrizo also on Fukumoto [21].

In addition, tree mortality in C-35 was notably relevant and likely due to the combination of both the multibad gall disorder and the clear iron chlorosis symptomatology on these trees. This rootstock is very sensitive to iron deficiency, which comprises its use in calcareous soils such as those found in Spanish citric areas. This behaviour has been reported in several studies [23,24,25,26]. CC is also catalogued as being sensitive to Fe chlorosis [27], but no visual symptoms were noted in our study. Therefore, more years should be spent investigating to reach more conclusions about this. Accordingly, ref. [23] found severer Fe deficiency symptoms in Navelina trees grafted onto C-35 than onto CC. In the same line, significant lower photosynthetic capacity and fluorescence Fv/Fm disorder were observed in a recent study in Clemenules mandarin grafted onto C-35 compared to CC and FA-5 [25].

3.3. Yield Parameters

Clemenrubi trees were 39.5% less productive than Orogros (average values between 2009 and 2016 of 24.8 kg tree⁻¹ and 41.0 kg tree⁻¹, respectively (

Table 2. Yield (individual data for 8 harvests from 2009 to 2016 and mean, in kg tree⁻¹), yield efficiency (in kg m⁻³) and alternate bearing index (ABI, in %) of the ‘Clemenrubi’ and ‘Orogrós’ Clementine mandarin trees on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005.

Variety	Rootstock	Yield a (kg Tree−1)									Yield Efficiency a,b (kg m−3 Year−1)	ABI a (%)
		2009	2010	2011	2012	2013	2014	2015	2016	Mean
Clemenrubi	CC	6.1 b	10.7 b	29.4 a	27.3 b	15.2 b	33.0 a	19.8 a	19.6 b	20.1 b	42.8 a	30.1 a
	C-35	10.2 a	19.1 a	39.8 a	41 a	25.9 a	33.8 a	19.6 a	25.3 ab	26.8 a	41.3 a	25.6 a
	FA-5	8.8 ab	14.3 ab	43.4 a	39.6 ab	25.9 a	35.5 a	20.0 a	33.1 a	27.6 a	44.0 a	26.2 a
	Mean	8.4 *	14.7 ns	37.5 ns	36.0 *	22.3 *	34.1 ns	19.8 ns	26.0 *	24.8 *	42.7 ns	27.3 ns
Orogros	CC	13.4 b	18.7 b	58.6 c	51.4 b	27.8 a	56.7 ab	32.8 a	39.7 a	37.4 b	35.0 a	29.5 a
	C-35	17.2 a	28.9 a	75.8 a	62.6 ab	34.4 a	49.6 b	34.8 a	45.7 a	43.6 a	33.9 a	26.5 a
	FA-5	16.5 a	24.0 a	67.9 b	64.9 a	35.2 a	63.2 a	24.2 a	39.1 a	41.9 ab	36.2 a	29.5 a
	Mean	15.7 *	23.9 **	67.4 ns	59.6 ns	32.5 ns	56.5 ns	30.6 ns	41.5 ns	41.0 ns	35.0 ns	28.5 ns

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and ns (non-significant). ^b Calculated as total cumulated yields per cubic meter of canopy volume.

). For each variety, the rootstock influence was slightly different. Although the lowest yields were obtained when grafted onto CC in Clemenrubi trees and was seen as a general trend, it was statistically significant in four of the eight harvests (2009, 2012, 2013, 2016), and no differences were observed between the Clemenrubi/C-35 and Clemenrubi/FA-5 combinations. The rootstock influence on Orogros yields showed a clear dependence on tree age. The trees on CC were statistically less productive in the first years of the experiment (2009 to 2012) than for the rest (between 13.1% and 21.6% lower than the mean yield in this variety in these years). This trend was lost at the end of the experiment as no statistical differences were found between rootstocks.

The yield efficiency in Clemenrubi was higher than in the Orogros trees (average 42.7 kg m⁻³ and 35.0 kg m⁻³, respectively), but no statistical differences between rootstocks were observed in any variety (Table 2). The ABI index was similar in both varieties (average 27.3% and 28.5%, respectively) with no differences between rootstocks (Table 2).

So, despite the apparent lower yields recorded in the trees on CC, its good yield efficiency (relation between accumulated yields and canopy volume) results (similar to C-35 and FA-5) means that it is impossible to rule this out in production terms. Good yields on CC have also been observed in Valencia sweet oranges on trees on Volkamer lemon, C. macrophylla and Carrizo citrange [28]. In the same line, although the ‘Navelate’ trees on the Volkamer lemon rootstock produced the highest yield every year, the mean was not statistically different from the yield of the trees on CC and C-13 [29], while around 50% yield reduction was recorded on Cleopatra mandarin and the FA-020326 rootstocks. In ‘Lane late’ navel oranges, good yield levels were observed on FA-31 and CC [30].

Moreover, the importance of the ABI results lies precisely in their constancy over time after eight harvests with no statistical differences between rootstocks. Alternate behaviour has been proven to be a serious problem that reduces the yield and the profitability of some citrus plantations. This is the case of some lemon varieties, such as Verna and V50 or Navel orange, for which the use of some FA-418 rootstocks was recommended to attenuate this problem [18,31]. The FA 5 rootstock did not behave as alternate when used on F49 and FE lemons [18] or on Lane Late orange [30].

3.4. Fruit Size

In the Clemenrubi trees, the combination on C-35 (top zone Figure 3B) gave the biggest statistically significant fruit (

Table 3. Fruit size parameters (weight, in g; and diameter, in mm) of the ‘Clemenrubi’ and ‘Orogrós’ Clementine mandarin trees on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005. The results are presented as the individual data for four harvests (2010 to 2013) and the mean.

Variety	Rootstock	Fruit Weight a (g)					Fruit Diameter a (mm)
		2010	2011	2012	2013	Mean	2010	2011	2012	2013	Mean
Clemenrubi	CC	81.4 a	76.1 b	90.2 ab	71.3 b	79.7 b	54.6 a	53.7 b	55.7 a	52.0 b	54.0 b
	C-35	85.0 a	92.7 a	92.2 a	88.8 a	89.7 a	55.0 a	56.0 a	55.3 a	55.4 a	55.4 a
	FA-5	82.1 a	87.0 a	85.6 b	73.5 b	82.1 b	54.7 a	55.3 ab	54.2 b	51.7 b	54.0 b
	Mean	82.8 ns	85.3 **	89.3 *	77.9 **	83.8 ***	54.8 ns	55.0 *	55.1 *	53.0 *	54.5 **
Orogros	CC	78.1 a	72.8 b	85.7 b	76.5 b	78.3 b	54.2 a	51.6 b	54.8 a	53.1 b	53.4 b
	C-35	82.1 a	81.8 a	90.8 a	87.8 a	85.6 a	54.9 a	53.6 a	55.5 a	55.7 a	54.9 a
	FA-5	78.5 a	86.6 a	90.0 a	84.1 ab	84.8 a	53.7 a	54.8 a	55.2 a	54.2 ab	54.5 a
	Mean	79.6 ns	80.4 **	88.8 *	82.8 ns	82.9 ***	54.3 ns	53.3 *	55.2 ns	54.3 ns	54.3 **

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) and ns (non-significant).

) in terms of both their weight (average 89.7 g, 12.6% and 9.3% higher than CC and FA-5, respectively) or diameter (average 55.4 mm, 2.6% and 2.6% higher than CC and FA-5, respectively). In the Orogros variety, both C-35 and FA-5 induced high fruit weight (average 85.2 ± 0.6, 8.1% higher than CC) and diameter (average 54.7 ± 0.3, 2.4% higher than CC) values, with no differences between them.

Despite the apparently influence of rootstock on fruit diameter, as a general rule no statistical effect was observed on the fruit shape index, calculated as the ratio between fruit diameter and height (

Table 4. Fruit shape (diameter/height) and peel thickness (in mm) of the fruit of the ‘Clemenrubi’ and ‘Orogros’ Clementine mandarins on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005. The results are presented as the individual data for four harvests (2010 to 2013) and the mean.

Variety	Rootstock	Fruit Shape a					Peel Thickness a (mm)
		2010	2011	2012	2013	Mean	2010	2011	2012	2013	Mean
Clemenrubi	CC	1.07 a	1.11 a	1.08 a	1.06 a	1.08 a	2.6 a	2.8 a	2.5 a	2.6 a	2.6 a
	C-35	1.08 a	1.11 a	1.06 b	1.07 a	1.08 a	2.3 b	2.4 b	2.0 b	2.4 ab	2.3 b
	FA-5	1.09 a	1.11 a	1.07 ab	1.05 a	1.08 a	2.4 b	2.4 b	2.1 b	2.0 b	2.2 b
	Mean	1.1 ns	1.1 ns	1.1 ns	1.1 ns	1.1 ns	2.4 *	2.5 *	2.2 *	2.3 *	2.4 ***
Orogros	CC	1.08 b	1.14 a	1.08 a	1.06 b	1.09 a	2.6 a	2.33 a	2.28 a	2.3 a	2.4 a
	C-35	1.11 a	1.11 a	1.07 a	1.11 a	1.10 a	2.4 a	2.11 ab	2.09 b	2.1 a	2.2 b
	FA-5	1.07 b	1.11 a	1.06 b	1.08 b	1.08 a	2.5 a	2.08 b	2.11 ab	1.9 a	2.1 b
	Mean	1.1 ns	1.1 ns	1.1 ns	1.1 *	1.1 ns	2.5 ns	2.2 ns	2.2 ns	2.1 ns	2.2 **

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) and ns (non-significant).

). Regarding peel thickness, the Clemenrubi/CC fruit presented the thickest fruit (average 2.6 mm, 11.5% and 15.4% higher than on C-35 and FA-5, respectively), and this result was also observed in all the harvest years. In Orogros, the average peel thickness in Orogros/CC also obtained the highest value (average 2.4 mm, 8.3% and 12.5% higher than on C-35 and FA-5, respectively), but no statistical differences were noted in each individual year.

Of the external fruit quality parameters, fruit size is one of the main parameters for fresh markets, and it is well-known that rootstock strongly influence it [15,16,28]. In general terms, C-35 induced bigger fruit in both varieties, while no differences were observed between the CC and FA-5 fruit. Ref. [32] reported that for ‘Marisol’ clementine and Gregoriou and Economides (1993) on ‘Shamouti’ orange that the trees on sour orange, CC, and Swingle citrumelo produced similar fruit in weight and size terms. When studying other citrus, Volkamer lemon, CC, and C-13 conferred ‘Navelate’ fruit the biggest sizes, while the lowest were for the Cleopatra mandarin and FA-020326 rootstocks [29]. The ‘Lane Late’ trees on FA-418 obtained the largest fruit diameter, followed by FA-31, Cleopatra mandarin, FA-030230 and FA-020324, while the fruit on FA-13 and CC were smaller [30]. However, in ‘Navelina’, the largest fruit were found on C-35 citrange, CC and Volkamer lemon [15], while the lowest size was for FA-13. Ref. [12] did not find any differences between the ‘W. Navel’ fruit on CC and Cleopatra mandarin. These findings demonstrate the high diversity of the results found in the bibliography, especially for this parameter, which can depend on not only plant material, but also on environmental conditions and cultural practices.

In any case, the average calibre obtained in our study was not a good result (around 54 mm on average, category III in Commission Implementing Regulation (EU) No. 543/2011) if compared to the biggest calibres from medium-season clementines (63–70 mm for Clementine de Nules, and 70–73 mm for Oroval, [33], or 75–80 mm for the Satsumas group, [5,33]. The ‘US Early Pride’ mandarin presented diameters of 68 mm, which are very interesting diameters considering its early-ripening behaviour [34]. All in all, despite the common trend for the fruit from early-season varieties having medium-sized diameters, the market shortage of mandarins at this time of year does not make this parameter as important as it would be halfway through the campaign.

3.5. Fruit Granulation Disorder

The percentage of fruit with dried vesicles was clearly dependent on rootstock (

Table 5. Fruit with dried juice vesicles and juice content (both in %) of the ‘Clemenrubí’ and ‘Orogrós’ Clementine mandarins on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005. The results are presented as the individual data for four harvests (2010 to 2013) and the mean.

Variety	Rootstock	Fruit with Dried Juice Vesicles a (%)					Juice a (%)
		2010	2011	2012	2013	Mean	2010	2011	2012	2013	Mean
Clemenrubi	CC	24.8 a	30.4 a	19.0 a	25.0 a	24.8 a	48.6 b	45.0 b	50.3 b	48.1 b	48.0 b
	C-35	1.4 b	2.8 b	1.0 b	0.0 b	1.3 b	52.2 a	52.1 ab	58.6 a	57.7 a	55.1 a
	FA-5	1.6 b	1.4 b	0.0 b	3.0 ab	1.5 b	52.7 a	54.7 a	57.2 a	56.7 a	55.3 a
	Mean	9.3 ***	11.5 ***	6.7 *	9.3 *	9.2 ***	51.2 *	50.6 ns	55.4 **	54.2 *	52.8 ***
Orogros	CC	19.2 a	7.0 a	23.0 a	28.0 a	19.3 a	50.6 c	53.5 b	52.7 b	47.1 b	51.0 b
	C-35	1.2 b	3.0 a	1.0 b	0.0 b	1.3 b	54.9 a	58.7 a	60.7 a	59.7 a	58.5 a
	FA-5	1.0 b	1.0 a	1.0 b	1.0 b	1.0 b	53.0 b	58.7 a	59.8 a	59.5 a	57.8 a
	Mean	7.1 *	3.7 ns	8.3 ***	9.7 **	7.2 ***	52.8 ***	57.0 **	57.7 ***	55.4 ***	55.8 ***

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) and ns (non-significant).

), and CC induced the highest values in both varieties (24.8% in Clementubi and 19.3% in Orogros), while disorder was very low in the fruit from the trees on the other two rootstocks (between 1.0% and 1.5%).

Ref. [10] described the rootstock influence on the proportion of fruit of ‘Lane late” sweet orange affected by granulation, as being CC and Citrumelo the rootstocks which generated the highest and lowest level of symptoms, respectively. The relatively high temperatures bloom, low fruit set, and larger fruit likely played an important role in the excessive granulation development on fruit [10]. Hence, changing cultural practices (i.e., fertilisation and irrigation) and using less vigorous rootstocks were proposed as good tools against this disorder, and vigour was ruled out because, according to the tree growth parameters, CC did not seem more vigorous than C-35 and FA-5. So, we can look at another factor.

As a rule, fruit increases in weight at night time and early in the morning when transpiration reaches its minimum [35]. Hydraulic conductance is considered one of the main factors that controls the movement of water through the soil–plant system [36] and will, therefore, strongly influence plant transpiration, the hydric state of leaves and, thus, the growth and physiological responses of plants.

From the bibliography, we found that CC is considered sensitive to hydric deficit [37] (Forner 1984), probably due to its low root hydraulic conductance [38,39]. The relatively low Kr of the roots in CC may influence leaf water deficit more intensely via a stronger regulation of the water flow from roots to leaves and can mean a shortage of the water transported to the aerial part. This is a notable disadvantage for the early varieties grafted onto this rootstock because the highest demands for water in fruit (from July to September) coincide with the time when transpiration for plants is higher. This would explain the larger number of fruits presenting dried vesicles. Moreover, under these conditions, plants would reduce not only transpiration and water loss via stomatal closure, but it also their photosynthetic capacity, as reflected by their smaller fruit and a tendency to smaller sized trees, at least in the Orogros variety. In this sense, [40] found a significant correlation (R = 0.63) between the gravimetrically determined plant transpiration rate and root hydraulic conductivity, apart from the leaf biomass in ungrafted rootstocks [38] and plant vigour [41,42].

3.6. Fruit Internal Quality

The lowest juice content was once again recorded in the fruit of Clemenrubi/CC and Orogros/CC (around 14.8% lower than in the other combinations), while no statistical differences were observed in the juice content between C-35 and FA-5 (Table 5). This trait located Clemenrubi/CC furthest to the left in Figure 3B, while Orogros/CC was placed in a central zone due to the big tree size typical of the variety.

As a general trend, TSS were similar between varieties and rootstocks (

Table 6. Total soluble solids (TSS) and total acids (TA) (both in %) of the fruit of the ‘Clemenrubí’ and ‘Orogrós’ Clementine mandarins on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005. The results are presented as the individual data for four harvests (2010 to 2013) and the mean.

Variety	Rootstock	TSS a (%)					TA a (%)
		2010	2011	2012	2013	Mean	2010	2011	2012	2013	Mean
Clemenrubi	CC	10.8 a	11.1 b	10.9 a	10.6 a	10.9 a	1.1 b	0.94 b	1.3 c	1.1 b	1.11 b
	C-35	10.7 a	11.3 ab	10.8 a	9.8 b	10.6 a	1.2 a	1.04 a	1.5 a	1.3 a	1.26 a
	FA-5	10.6 a	11.8 a	10.7 a	10.4 a	10.9 a	1.3 a	0.98 ab	1.4 b	1.3 a	1.23 a
	Mean	10.7 ns	11.4 ns	10.8 ns	10.3 *	10.8 ns	1.2 *	1.0 ns	1.4 ns	1.2 *	1.2 ***
Orogros	CC	11.2 a	10.8 a	11.3 a	9.9 a	10.8 a	1.3 a	1.0 a	1.4 a	1.1 c	1.19 b
	C-35	10.8 ab	10.7 a	11.0 a	10.1 a	10.6 a	1.2 a	1.0 a	1.5 a	1.4 a	1.28 a
	FA-5	10.5 b	10.9 a	11.2 a	10.0 a	10.6 a	1.4 a	0.9 a	1.5 a	1.2 b	1.24 ab
	Mean	10.8 *	10.8 ns	11.2 ns	10.0 ns	10.7 ns	1.3 ns	1.0 ns	1.5 ns	1.2 ***	1.2 *

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.001 (***) and ns (non-significant).

), with an average value of 10.8 ± 0.2 and 10.7 ± 0.1 for Clemenrubi and Orogros, respectively. The acidity index was statistically lower in the fruit from the trees grafted onto the CC rootstock. The TA in the Clemenrubi/CC fruit was 0.14% lower than the mean value of this variety in the other two combinations (1.25 ± 0.02%). Albeit lower, the TA in the Orogros/CC fruit was also statistically lower (average 0.07%) than the mean value of this variety in the other two combinations (1.26 ± 0.03%).

The RI also depended on rootstock (

Table 7. The ripening index (RI, TSS/TA) and colour index (CI) of the fruit of the ‘Clemenrubi’ and ‘Orogros’ Clementine mandarins on three rootstocks (Carrizo citrange, CC, control; C-35 citrange, C-35 and Forner-Alcaide 5, FA-5) planted in July 2005. The results are presented as the individual data for four harvests (2010 to 2013) and the mean.

Variety	Rootstock	RI a					CI a
		2010	2011	2012	2013	Mean	2010	2011	2012	2013	Mean
Clemenrubi	CC	9.8 a	11.8 b	8.5 a	9.4 a	9.9 a	−11.1 a	−4.1 a	−8.2 a	−10.1 a	−8.4 a
	C-35	8.6 b	10.8 ab	7.4 b	7.6 b	8.6 c	−11.8 a	−7.6 a	−9.1 a	−15.0 b	−10.9 b
	FA-5	8.4 b	12.1 a	7.8 b	8.0 b	9.1 b	−10.6 a	−4.4 a	−9.6 a	−13.8 b	−9.6 ab
	Mean	8.9 *	11.6 ns	7.9 *	8.3 **	9.2 ***	−11.2 ns	−5.4 ns	−9.0 ns	−13.0 *	−9.6 **
Orogros	CC	8.6 a	10.7 ab	8.2 a	9.4 a	9.2 a	−11.4 a	−12.0 a	−8.8 a	−10.9 a	−10.8 a
	C-35	8.8 a	10.4 b	7.4 a	7.4 c	8.5 b	−15.1 b	−10.4 a	−10.3 b	−14.6 b	−12.6 ab
	FA-5	7.7 a	11.5 a	7.7 a	8.3 b	8.8 ab	−16.5 b	−10.8 a	−10.7 b	−16.0 b	−13.5 b
	Mean	8.4 ns	10.9 ns	7.8 ns	8.4 **	8.8 *	−14.3 *	−11.1 ns	−9.9 *	−13.8 *	−10.5 **

^a Mean separation in columns by the Duncan test (p < 0.05). For the same variety, different letters denote statistical differences between rootstocks. Levels of significance are represented by p < 0.05 (*), p < 0.01 (**) and p < 0.001 (***) and ns (non-significant).

). As a general trend, the fruit of both varieties grafted onto CC had the highest values in not only most individual harvests, but also in the mean of 4 years. The average value of 4 years in Clemenrubi/CC was 10.6% higher than on C-35 and FA-5 (8.9 ± 0.35). In Orogros, the average value on CC was 6.0% higher than the mean value on C-35 and FA-5 (8.7 ± 0.2).

Regarding the CI (Table 7), the fruit from the CC combinations had a statistically significant higher CI value than for the other rootstocks. The average CI values for the Clemenrubi/CC and Orogros/CC fruit were 17.3% and 12.8% higher than the mean value of the other two rootstocks in Clemenrubi and Orogros (−12.6 ± 0.8 and −15.0 ± 0.7, respectively).

Two good indicators of the fruit maturity state are the RI, which reflects the balance between TSS and TA in juice, and the peel CI. The fruit on FA-5, and especially those on C-35, obtained lower RI values, which indicates a certain delayed maturity process of these fruit (right zone Figure 3B). With no statistical differences in the TSS results, this is a consequence of their higher acidity level. Rootstock influences important quality parameters related to not only external traits, but also to internal characteristics (size, peel thickness and content, juice and sugar contents, etc.), as previously reported [16,17,29,30,43]. For the marketing standard for citrus fruit, Commission Implementing Regulation (EU) No. 543/2011) sets out mandarin maturity requirements, such as 40% minimum juice content, a TSS/TA ratio of 7.0:1 and the variety’s typical colour seen on at least one third of the fruit surface. Thus, according to this regulation, all the scion/stock combinations herein evaluated obtained appropriate values. Good quality results (close to our RI, TSS and TA data) were described in the ‘US Early Pride’ mandarin, the early-maturing mandarin hybrid obtained in Florida which avoided the seed production of its predecessor ‘Fallglo’ mandarin [34].

This is not a minor point for varietal breeding programmes because in a similar study carried out on 15 Clementine mandarin selections and three hybrids in Brazil, only a few varieties developed adequate minimum standards for all the parameters [33,44]. Similarly, according to another study on 46 tangerines and hybrids, only two and four were selected for their good attributes and precocity or medium season habit, respectively [33].

Moreover, narrowly exceeding the required minimum CI is not such a marked limitation for harvesting as it was in the past because important advances in degreening techniques have taken place [45]. Even so, a high CI can be a disadvantage if fruit have obtained good internal maturity values because this would coincide with the maximum activity of the Mediterranean fruit fly, Ceratitis capitata (Wiedemann), a major threat for these early-season varieties. So, the lower CI in the fruit on C-35 and FA-5 compared to CC suggests their initial advantage for these combinations.

3.7. PCA Analysis

The PCA analysis (Figure 3) and those eigenvalues above 1 reflected (Table S1) a different pattern in the correlation of scion/rootstock combinations. Two principal components were determined, which described around 89.5% of the variability between combinations.

The first component principally correlated with juice quality and tree growth parameters (Figure 3A). All the correlations were moderate, and the strongest positive relations were observed with tree height and canopy diameter and volume, in addition to fruit content. Negative correlations were found with internal fruit quality parameters (dried vesicles, TSS, RI, and CI). Therefore, the biggest trees with the poorest fruit quality were placed to the right of the plot, while the smallest trees and best fruit quality were placed to the right of the plot. In line with this, when analysing the second component, the highest positive correlations were with tree mortality, fruit size parameters and yield efficiency, while negative correlations were found for ABI percentage. So, the combinations which presented big fruits but also strong mortality incidence were placed in the upper part of the plot, while the trees with the lowest ABI values were placed in a lower position. The projection on the PCA plot (Figure 3B) showed how combinations spread widely over the area. In general, Orogros trees were located on the low part of the PCA due to their big trees and low yield efficiency. Regarding rootstocks, trees grafted on C-35 and FA-5 remained together while CC trees were placed separately. This distribution is also supported by the results from the dendogram chart and the distance found between scion/rootstocks combinations (Figure S1).

4. Conclusions

The present study investigated the effects of three rootstocks on yield and some fruit quality characteristics of two very early-season mandarins, Clemenrubi and Orogros, as an initial study to characterise their alternative for moving forward the availability of this fruit on the market calendar.

The most remarkable variety effect was less Clemenrubi tree development and their lower yields.

The lowest yields and smallest fruit size (weight and diameter) were obtained on CC, while no differences were found between C-35 and FA-5 trees. Yield efficiency and ABI were not affected by rootstock. All the scion/stock combinations showed a good juice internal quality parameter and met the minimum marketing requirements.

C-35 was also highlighted for delaying fruit maturity (lower RI and CI indices), but no significant differences in TSS were obtained compared to the other rootstocks. So, this fruit is also suitable for fresh markets.

CC, control rootstock, also presented a marked granulation disorder. Although all the scion/combinations showed multibud galls, the highest tree mortality was observed for the trees grafted onto C-35, likely because of its sensitivity to Fe chlorosis.

As a whole, all the rootstocks for Clemenrubi and Orogros mandarins gave good fruit internal quality for fresh fruit markets in the eastern Mediterranean region.

It is worth mentioning that long-term studies are needed to determine the exact effects of multibud galls on the tree survival and granulation problem on the CC fruit.