Abstract
Timing of flowering is a fundamental developmental transition that has great ecological and agricultural importance. For perennial plants, seasonal timing of bloom and anthesis, which is the ultimate stage of flowering, can be determined by the net effect of several preceding developmental steps: seasonal timing of floral initiation, rate and extent of floral development before winter dormancy, duration of dormancy, and rate of floral development after release from dormancy. In the domestic apple (Malus ×domestica), fruit production has generally favored cultivars that bloom relatively early in the season. However, floral tissues are easily damaged by freezing temperatures, and freeze injury is especially problematic in years when abnormally warm temperatures in early spring lead to rapid floral development. To facilitate identification of genes/alleles that govern bloom time, and that could add versatility to production systems for apple, we evaluated seasonal bloom time for accessions of M. ×domestica, wild apple species (Malus sp.), and Malus hybrids maintained in a large germplasm diversity collection.
One of the most important plant traits, in terms of both ecology and agricultural production, is seasonal timing of flowering. Seasonal flowering is driven by the interaction of endogenous genetic pathways with environmental cues such as photoperiod, light intensity and wavelength, and temperature (Amasino, 2009; Andrés and Coupland, 2012; Lang, 1965). In this way, flowers are produced at a time of the season that maximizes the potential of a given plant to produce abundant fruit and seed. For self-incompatible plants, precise control of flowering in a population is especially important to ensure cross-pollination, and for plants that rely on biotic pollination, seasonal control is important to ensure flowering at a time when specific pollinators are present and active.
For many woody perennials growing in temperate climates, including several major horticultural fruit species, flowering spans two growing seasons interrupted by an extended period of winter dormancy. Thus, flowers are differentiated on the shoot meristem during the first season and develop incompletely before the onset of dormancy. During the subsequent spring, flowers complete development, culminating in bloom and anthesis. When plants have perceived an appropriate period of winter temperatures (chilling requirement or endodormancy), this further development is dependent only on environment including warmer temperatures (ecodormancy) (Lang et al., 1987). In many seasons, extensive development associated with abnormally warm temperatures after completion of endodormancy leads to susceptibility to frost damage. A striking example was seen in 2012 throughout the midwestern United States, where prolonged high temperatures in early spring led to bloom of many horticultural crops more than 1 month before normal (Fig. 1). This period was followed by a series of hard freezes, which led to widespread and extensive losses for many of these including peach (Prunus persica), cherry (Prunus cerasus and Prunus avium), and apple [U.S. Department of Agriculture (USDA), 2012]. For apple, which can be strongly biennial for flowering and fruit production, this degree of loss may lead to overcropping the subsequent year, a problem that must be countered by expensive chemical or manual thinning of flowers or fruit.
Average daily temperatures and winter chill unit accumulation for Geneva, NY, in 2009–10 and 2011–12. Chill units were determined by the method of Hauagge and Cummins (1991).
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Average daily temperatures and winter chill unit accumulation for Geneva, NY, in 2009–10 and 2011–12. Chill units were determined by the method of Hauagge and Cummins (1991).
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Average daily temperatures and winter chill unit accumulation for Geneva, NY, in 2009–10 and 2011–12. Chill units were determined by the method of Hauagge and Cummins (1991).
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
If climate change leads to more variability in early-season temperatures, then crop losses such as that seen in 2012 will become increasingly common. An important strategy to sustain current production, therefore, will be to develop cultivars that are less prone to early bloom. This will in turn require the integration of phenological and genetic studies to identify genes/alleles that control aspects of flowering. Timing of spring bloom is expected to be under complicated genetic control, because it can be influenced by several factors including seasonal timing of floral initiation, rate and extent of floral development before winter dormancy, duration of endodormancy, and rate of floral development after release from dormancy (Bubán, 1996; Hänninen and Tanino, 2011; Soltész, 1996). However, the genetic control of such individual traits is mostly unknown, and general diversity in these traits among species and domesticated cultivars has generally not been explored.
To hasten the identification of the potential range of genes/alleles that govern timing of bloom, and that could add versatility to production systems for apple, we sought to establish the extent of natural variation seen in bloom time throughout the genus Malus. An extensive collection of Malus species, hybrids, and domestic apple cultivars is maintained at the USDA-Agricultural Research Service (ARS) Plant Genetic Resources Unit in Geneva, NY. A core subset considered to represent the diversity of the entire collection is also maintained at the Geneva site, allowing for efficient evaluation of important traits relevant to industrial production (Hokanson et al., 2001; Kresovich et al., 1995). The objectives of this study were to 1) document relative flower development within the core germplasm collection at multiple points during the typical period of bloom; and 2) identify those accessions with the most extreme (early or late) bloom time within the core and extended collection.
Materials and Methods
This study used the extensive Malus germplasm collection maintained at the USDA-ARS Plant Genetic Resources Unit in Geneva, NY. Plantings were managed in accordance with routine commercial practices for weeds, irrigation, nutrition, and microbial and insect pests. Trees were maintained as 16-year-old grafts of clonal material on rootstock ‘P.22’, 5-year-old grafts on ‘BUD 9’ (B.9), 8-year-old grafts on ‘Malling 7’ (M.7), and/or 10-year-old grafts on ‘EMLA 7’ (E.7). Evaluations were carried out on 21 Apr. 2010, 1 May 2010, 15 May 2010, and 2 May 2012. Floral development stage was evaluated for five representative inflorescences (flower clusters) for each individual using the numerical assessment scheme described in the text and an average value was calculated. Imprecision in evaluation was assessed by comparing ratings, taken on the same day by the same evaluator, for 171 accessions in the B9 planting. Most (61%) differences in measurements were 0.2 units or less, and 92% of differences in measurements were 1 unit or less. Those with measurement differences of 1 or greater were nearly all at a relatively early developmental stage for which developmental gradation was less obvious. Accumulated chilling hours as shown in Figure 1 were determined by the method of Hauagge and Cummins (1991), beginning on 1 Nov. Data from this study have been submitted to the USDA Germplasm Resources Information Network and are available by request from the corresponding author.
Results
Diversity in floral development among apple genotypes.
As an initial assessment of diversity in floral development, we catalogued developmental stage for 189 accessions of the core diversity collection on three dates in Spring 2010 (Table 1). An additional 50 individuals in this core collection either showed no flowering or exhibited only a few inflorescences and were not evaluated. Lack of flowering in apple can result from the natural phenomenon of biennial bearing. The evaluation dates were chosen in accordance with early (21 Apr.), intermediate (1 May), and late (15 May) bloom timing for the population in general in 2010. A numerical assessment scheme was developed (Fig. 2; Table 2) that is less sensitive to inflorescence and flower structural variation among species than a traditional apple bloom progression scale often used for M. ×domestica (Chapman and Catlin, 1976). For each of 68 accessions that were represented by one or more individuals, measurements varied only slightly with a median difference between replicates of 0.1 units and 90% of replicates within 0.3 units (data not shown).
Malus specimens and accessions evaluated in this study in 2010.
Stages of Malus floral development defined in this study.
Stages of floral development in apple as designated in this study. (A) At the earliest stage (Stage 0), plants had just broken dormancy with vegetative or mixed vegetative/floral shoots apparent. Inflorescences were visible only upon macroscopic dissection of shoots. (B) At Stage 1, the inflorescence was macroscopically visible within the shoot without dissection (Malus ×domestica). (C–D) By Stage 2, individual flowers were easily apparent as a tight cluster [e.g., in (C) M. ×domestica] or loose cluster [e.g., (D) M. fusca]. At this stage, in (C) M. ×domestica, petals were visible in the dominant “king” flower. (E–F) By Stage 3, most flowers within the inflorescence showed petals [(E) M. ×domestica, (F) M. ioensis]. (G–H) By Stage 4, petals were apparent in all flowers [(G) M. ×domestica, (H) M. fusca]. (I) Petals became dominant over sepals on the bud in Stage 5 (M. ioensis). (J) In Stage 6, petals of one or more flowers failed to totally enclose the floral bud. (K) One or more flowers was fully open at Stage 7. (L) At Stage 8, all flowers in an inflorescence were open.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Stages of floral development in apple as designated in this study. (A) At the earliest stage (Stage 0), plants had just broken dormancy with vegetative or mixed vegetative/floral shoots apparent. Inflorescences were visible only upon macroscopic dissection of shoots. (B) At Stage 1, the inflorescence was macroscopically visible within the shoot without dissection (Malus ×domestica). (C–D) By Stage 2, individual flowers were easily apparent as a tight cluster [e.g., in (C) M. ×domestica] or loose cluster [e.g., (D) M. fusca]. At this stage, in (C) M. ×domestica, petals were visible in the dominant “king” flower. (E–F) By Stage 3, most flowers within the inflorescence showed petals [(E) M. ×domestica, (F) M. ioensis]. (G–H) By Stage 4, petals were apparent in all flowers [(G) M. ×domestica, (H) M. fusca]. (I) Petals became dominant over sepals on the bud in Stage 5 (M. ioensis). (J) In Stage 6, petals of one or more flowers failed to totally enclose the floral bud. (K) One or more flowers was fully open at Stage 7. (L) At Stage 8, all flowers in an inflorescence were open.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Stages of floral development in apple as designated in this study. (A) At the earliest stage (Stage 0), plants had just broken dormancy with vegetative or mixed vegetative/floral shoots apparent. Inflorescences were visible only upon macroscopic dissection of shoots. (B) At Stage 1, the inflorescence was macroscopically visible within the shoot without dissection (Malus ×domestica). (C–D) By Stage 2, individual flowers were easily apparent as a tight cluster [e.g., in (C) M. ×domestica] or loose cluster [e.g., (D) M. fusca]. At this stage, in (C) M. ×domestica, petals were visible in the dominant “king” flower. (E–F) By Stage 3, most flowers within the inflorescence showed petals [(E) M. ×domestica, (F) M. ioensis]. (G–H) By Stage 4, petals were apparent in all flowers [(G) M. ×domestica, (H) M. fusca]. (I) Petals became dominant over sepals on the bud in Stage 5 (M. ioensis). (J) In Stage 6, petals of one or more flowers failed to totally enclose the floral bud. (K) One or more flowers was fully open at Stage 7. (L) At Stage 8, all flowers in an inflorescence were open.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
The early evaluation date revealed a tremendous range in developmental stage among all accessions. The least developed among the 189 accession groups was a representative of M. angustifolia (southern crab apple), which had just broken dormancy and did not exhibit a visible inflorescence without dissection (Stage 0). Conversely, the most advanced was an accession of M. sylvestris (european crab apple) with all flowers open (Stage 8) (Fig. 3).
Stage of floral development for representatives of Malus species groups. Chart depicts individual observations for 189 accessions of the core Malus diversity collection on three dates in 2010 representing early [21 Apr. (open)], mid [1 May (gray)], and late [15 May (black)] development for the population in general. The number of accessions evaluated within each species group is indicated by size of the circle; here the smallest circle represents one accession. Medians of observations are shown as a short horizontal bar. Data represent ratings of floral development where 0 = buds that have just broken dormancy and 9 = nearly all flowers have shed petals, as described in Table 2 and Figure 2.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Stage of floral development for representatives of Malus species groups. Chart depicts individual observations for 189 accessions of the core Malus diversity collection on three dates in 2010 representing early [21 Apr. (open)], mid [1 May (gray)], and late [15 May (black)] development for the population in general. The number of accessions evaluated within each species group is indicated by size of the circle; here the smallest circle represents one accession. Medians of observations are shown as a short horizontal bar. Data represent ratings of floral development where 0 = buds that have just broken dormancy and 9 = nearly all flowers have shed petals, as described in Table 2 and Figure 2.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
Stage of floral development for representatives of Malus species groups. Chart depicts individual observations for 189 accessions of the core Malus diversity collection on three dates in 2010 representing early [21 Apr. (open)], mid [1 May (gray)], and late [15 May (black)] development for the population in general. The number of accessions evaluated within each species group is indicated by size of the circle; here the smallest circle represents one accession. Medians of observations are shown as a short horizontal bar. Data represent ratings of floral development where 0 = buds that have just broken dormancy and 9 = nearly all flowers have shed petals, as described in Table 2 and Figure 2.
Citation: Journal of the American Society for Horticultural Science J. Amer. Soc. Hort. Sci. 138, 5; 10.21273/JASHS.138.5.367
The early evaluation date also revealed an extensive developmental range within species groups. For example, for the species M. hupehensis, accessions ranged from just after Stage 1, with the inflorescence just visible within the new shoot, to after Stage 6, when many flowers had begun to open (Fig. 3). A similar range was seen in M. baccata, where accessions ranged from just after Stage 2 (petals becoming visible in buds) to nearly Stage 8. Accessions that were included in the large M. ×domestica and Malus hybrid groups were similarly broadly distributed along developmental stage (Fig. 3). The M. ×domestica cultivar that had undergone the least development was ‘Cox's Orange Pippin’, traditionally considered as late-blooming, whereas the named cultivar that had advanced furthest was ‘Novosibirski Sweet’, originally sourced from the former Soviet Union (USDA, 2013).
The intermediate evaluation date revealed the greatest range in floral development between species groups (Fig. 3). Some or all representatives of M. prattii, M. angustifolia, M. florentina, M. ombrophila, M. rockii, and M. transitoria exhibited extremely compact inflorescences with no petals visible (before Stage 2), whereas for some or all representatives of M. sieversii, M. brevipes, M. orientalis, and M. sylvestris, most flowers in an inflorescence had opened (Stage 7.5) (Fig. 3).
These data from the core collection showed that bloom development could be strongly variable both within a species group and between species groups. To focus analysis on bloom, we queried the entire Malus collection for specimens that represented the extreme range in bloom time. This effort considered nearly 2700 individuals representing nearly 1800 distinct accessions that were available for evaluation and were flowering in the year of this study (Table 1).
We first identified individuals that were most advanced at the time of the early assessment. We identified 121 individuals representing 76 accessions that had reached Stage 7 (one flower fully opened) and 46 individuals representing 31 accessions that had reached Stage 7.3 (50% open flowers). We defined these as “early blooming” and “extreme early blooming,” respectively (Tables 3 and 4). In addition to M. sylvestris (described previously), individual accessions of M. baccata and M. orthocarpa had reached Stage 8. The M. ×domestica cultivar that was most advanced was ‘Shinibalt 1’, originally sourced from Pakistan, which had reached Stage 7.8 (90% of flowers open) (Tables 3 and 4). When considering those accessions that had reached Stage 7, M. ×domestica was underrepresented, whereas M. kirghisorum, M. orientalis, M. sieversii, and M. sylvestris were overrepresented (P ≤ 0.0015, Fisher’s exact test) relative to the collection population in general (Table 5).
Extreme early blooming accessionsz identified in the USDA–Geneva Malus collection.
Early blooming Malus ×domestica accessions identified in the USDA–Geneva Malus collection.z
Early blooming species groups identified in the USDA–Geneva Malus collection.
We then identified individuals that had not yet reached Stage 8 by the time of the late assessment. Here, we found 68 specimens representing 49 accessions that were at or less advanced than Stage 8 and 14 specimens representing eight accessions that were at or less advanced than Stage 7. We defined these as “late blooming” and “extreme late blooming,” respectively (Table 6). The developmentally least advanced accessions, which were at or before Stage 6, were of M. ombrophila, M. florentina, and M. angustifolia (Table 6). The developmentally least advanced accession of M. ×domestica was ‘Koningszuur’, a cultivar developed in The Netherlands, possibly incorporating M. sylvestris (USDA, 2013). Among late-blooming accessions, M. ×domestica and Malus hybrids were underrepresented, whereas M. angustifolia, M. coronaria, M. florentina, M. ioensis, and M. yunnanensis were overrepresented [P ≤ 0.001 Fisher’s exact test (Table 7 and data not shown)].
Extreme late-blooming accessions and late-blooming Malus ×domestica accession identified in the USDA–Geneva Malus collection.z
Late-blooming species groups identified in the USDA–Geneva Malus collection.
Year-to-year variation in bloom stage.
To assess year-to-year variability in floral development in relative terms among genotypes, we recorded developmental stage for 172 accessions in the B.9 planting in 2012 on a date (2 May) similar to that for the intermediate evaluation in 2010 (1 May). As expected as a result of warmer than typical early spring conditions in 2012 (Fig. 1), most accessions (118) were more advanced in 2012 relative to 2010 on this date (Table 8 and data not shown). Interestingly, a small subset of accessions (38) was less advanced on this date in 2012. Analysis of the species composition of this subset (Table 8) revealed that M. ×domestica accessions were significantly overrepresented, representing 55% of this subset but only 28% of the analyzed population [χ2 P value 0.007 (data not shown)]. Conversely, M. ×domestica accessions were underrepresented among those that showed the most advancement of development relative to 2010 [P = 0.022 (data not shown)]. Accordingly, relative bloom stage between 2010 and 2012 was well correlated for non-domestica accessions (Pearson's correlation 0.8826, n = 123) but only poorly correlated for domestica accessions [0.4478, n = 49 (data not shown)]. This may suggest that many domestica accessions may be recalcitrant to excessive promotion of bloom by early spring warmth. Alternatively or in addition, this effect could result from variation in chill requirement among accessions and the dissimilar winter conditions experienced in 2009–10 vs. 2011–12. We analyzed our data with respect to a study of variation in chill requirement among selected representatives of the same germplasm collection carried out by Hauagge and Cummins (1991). For the 43 accessions that were analyzed in both studies, we found a negative correlation between chill requirement and developmental stage, which was stronger in 2012 (Pearson’s correlation –0.5696) than in 2010 (–0.3185) (data not shown). The exceptional low chill-requiring cultivar Anna (Hauagge and Cummins, 1991) was among those accessions showing the greatest advancement in 2012 relative to 2010. Whereas these observations would not be inconsistent with delayed development response for higher chill-requiring accessions in 2012, calculated chill unit accumulation in the relatively warm winter of 2011–12 surpassed that of the more typical winter of 2009–10 as a result of extended periods of temperatures just above freezing in 2011–12 (Fig. 1). Further speculation on genetic response to variation in seasonal temperature regimes will be facilitated by a morphological or genetic marker for the completion of endodormancy (Soltész, 1996).
Accessions in the USDA–Geneva Malus collection that showed less advanced or more advanced floral development in 2012 relative to 2010.z
Discussion
Timing of spring bloom and its response to seasonal temperatures is a critical trait for production, and development of new cultivars that bloom later in the season should be an increasing priority. To facilitate identification of genetic allele(s) that control this trait, we thus initiated assessment of bloom stage in an extensive diversity collection. Our results reveal substantial range of bloom stage both between and within species. Whereas a wide range was not unexpected among the many species of apple, the range seen within some species [e.g., M. hupehensis (Fig. 3)] was unanticipated, and the expression of such diversity within a species suggests this trait may be controlled by one or a small number of alleles. Simple genetic control of flowering has been seen in Arabidopsis thaliana in nature, in which a single gene controls the striking difference in flowering habit between annual and winter-annual accessions (Gazzani et al., 2003).
The substantial variation in bloom stage within M. ×domestica cultivars was expected given the heterogeneous origins of domestic apple (Cornille et al., 2012). This variation did not seem to be conditioned by tree architecture, because the rootstocks, training systems, and management regimes were similar. Although flowers in lateral positions on long shoots can be delayed in opening relative to those arising on spurs (Soltész, 1996), in this study, we did not find obvious differences in developmental stage strictly associated with shoot-dominated vs. spur-dominated flowering architectures.
Malus ×domestica cultivars were, in general, earlier to bloom than most Malus species. Commercial apple production has generally favored cultivars that bloom relatively early in the season, anecdotally to produce marketable fruit as early as possible in the season, and to ensure fruit maturity in short growing seasons (Soltész, 1996). In this study, we did not find a significant relationship between bloom stage and fruit maturity period for the 210 accessions of M. ×domestica for which seasonal period of fruit maturity was previously recorded (USDA, 2013) [Pearson's correlation –0.2769 (data not shown)]. This lack of association is consistent with results of the majority of studies reviewed by Soltész (1996). Other factors that may be more critical for early maturity, including rapid rate of fruit development and early initiation of fruit ripening, are thus not incompatible with late bloom as breeding goals.
Timing of bloom may be influenced by the sum effect of several preceding developmental steps, including seasonal timing of floral initiation, rate and extent of floral development before winter dormancy, duration of dormancy, and rate of floral development after release from dormancy. Whereas timing of bloom is an easily accessible trait to evaluate, the remaining developmental steps are more cryptic. Initiation, for example, takes place at the shoot apex, which is both microscopic and encased in vegetative tissues within the bud. Extent of endodormancy can only be evaluated destructively through collection of shoots in the field and incubation under warm temperatures. The effect of environment on development is expected to be complex; for example, whereas heat accumulation after endodormancy release promotes floral development, lack of sufficient chilling hours over winter can delay bloom (Chandler, 1942; Hänninen and Tanino, 2011). The identification of species and M. ×domestica cultivars that exhibit extremes in bloom stage provides an entry point for more advanced and focused analyses of the genetic and environmental factors contributing to bloom time.
Our study population comprised much of the known cataloged germplasm for Malus species and M. ×domestica cultivars and provides a valuable resource for practical breeding and association mapping. That variation in development among species was wider than that within M. ×domestica suggests potential for genetic improvement of M. ×domestica through wide (interspecific) crosses. We noted that the three accessions of the Malus species native to the study area, M. ioensis, were markedly later blooming than nearly all M. ×domestica accessions; in 2010, none of the flowers of M. ioensis had opened at the time when most M. ×domestica accessions were nearly in full bloom (Fig. 3). Although most M. ×domestica accessions at the study site suffered complete loss of crop in 2012, M. ioensis escaped frost damage and produced abundant fruit. M. ioensis is easily intercrossed with M. ×domestica and thus is a potential source of wild germplasm for introduction of late-blooming genetic alleles. In contrast, M. sieversii, a large-fruited species from central Asia believed to contribute much of the M. ×domestica genome (Cornille et al., 2012), is currently being used in wide crosses to introduce disease resistance and novel production traits. However, nearly all accessions of M. sieversii studied here bloom early, even relative to M. ×domestica (Fig. 3), and the responsible genetic alleles will need to be eliminated to produce an acceptable cultivar for frost-prone regions.
Literature Cited
Amasino, R. 2009 Floral induction and monocarpic versus polycarpic life histories Genome Biol. 10 228
Andrés, F. & Coupland, G. 2012 The genetic basis of flowering responses to seasonal cues Nat. Rev. Genet. 13 627 639
Bubán, T. 1996 Flower development and formation of sexual organs, p. 3–54. In: Nyéki, J. and M. Soltész (eds.). Floral biology of temperate zone fruit trees and small fruits. Akadémiai Kiadó, Budapest, Hungary
Chandler, W.H. 1942 Deciduous orchards. Lea and Febiger, Philadelphia, PA
Chapman, P.J. & Catlin, G.A. 1976 Growth stages in fruit trees—From dormant to fruit set Plant Sci. Entomol. (Geneva, Switzerland) 11 58
Cornille, A., Gladieux, P., Smulders, M.J.M., Roldán-Ruiz, I., Laurens, F., Le Cam, B., Nersesyan, A., Clavel, J., Olonova, M., Feugey, L., Gabrielyan, I., Zhang, X., Tenaillon, M.I. & Giraud, T. 2012 New insight into the history of domesticated apple: Secondary contribution of the european wild apple to the genome of cultivated varieties PLoS Genet. 8 E1002703
Gazzani, S., Gendall, A.R., Lister, C. & Dean, C. 2003 Analysis of the molecular basis of flowering time variation in Arabidopsis accessions Plant Physiol. 132 1107 1114
Hänninen, H. & Tanino, K. 2011 Tree seasonality in a warming climate Trends Plant Sci. 16 412 416
Hauagge, R. & Cummins, J.N. 1991 Phenotypic variation and length of bud dormancy in apple cultivars and related Malus species J. Amer. Soc. Hort. Sci. 116 100 106
Hokanson, S., Lamboy, W., Szewc-McFadden, A. & McFerson, J. 2001 Microsatellite (SSR) variation in a collection of Malus (apple) species and hybrids Euphytica 118 281 294
Kresovich, S., Lamboy, W., McFerson, J. & Forsline, P. 1995 Integrating different types of information to develop core collections, with particular reference to Brassica oleracea and Malus × domestica, p. 147–154. In: Hodgkin, T., A.H.D. Brown, T.J.L. Hintum, and E.A.V. Morales (eds.). Core collections of plant genetic resources. Wiley, Chichester, UK
Lang, A. 1965 Physiology of flower initiation, p. 1380–1536. In: Lang, A. (ed.). Encyclopedia of plant physiology. Vol. XV/1. Springer-Verlag, Berlin, Germany
Lang, G.A., Early, J.D., Martin, G.C. & Darnell, R.L. 1987 Endo-, para-, and ecodormancy: Physiological terminology and classification for dormancy research HortScience 22 371 377
Soltész, M. 1996 Flowering, p. 80–131. In: Nyéki, J. and M. Soltész (eds.). Floral biology of temperate zone fruit trees and small fruits. Akadémiai Kiadó, Budapest, Hungary
U.S. Department of Agriculture 2012 Crop production. 6 Mar. 2013. <http://www.nass.usda.gov/Publications/Todays_Reports/reports/crop0810.pdf>
U.S. Department of Agriculture 2013 Apple. 6 Mar. 2013. <http://www.ars-grin.gov/cgi-bin/npgs/html/crop.pl?115>
