Flower bud development of almond cultivars based on three different methods

Three twigs with 1-year-old laterals were collected weekly from each accession every year. In the laboratory, 10 flower buds per accession were selected randomly. The anthers were removed by a tweezer, stained with carminic acetic acid and squash preparations were made for microscopic studies. The microspore development stage (archesporium, string, pollen mother cells [PMC], tetrad cells, microspores and pollen cells) of each microspore was recorded. The proportion of each stage was calculated by accessions and sampling dates. On the basis of weekly data, we estimated when 50% of the stages occurred and this calendar date was regarded as the transmission date from one stage to another. An accession having at least 50% of its flower buds in the string stage is regarded as reaching the end of endodormancy.

Pistils of the 10 flower buds per accession that were used in microsporogenesis studies were examined. The length of the pistil was recorded on the microscope slide equipped with a stage micrometer (Carl Zeiss, Germany), with an accuracy of 0.1 mm. The resumption of pistil growth after being constant was considered as endodormancy release.

Three twigs with 40–50 flower buds of each accession were collected weekly, from the beginning of November until the end of February to the beginning of March. One twig with 10–20 flower buds was taken as one replication. The twigs were transferred to the laboratory and immediately placed in water with their basal end cut in 1-L containers with 0.5 L of water and kept at room temperature with natural photoperiod reflected through the window. After 10 days, the number of open flowers was recorded. The date when accessions had 50% of open flowers was regarded as the end of endodormancy.

To evaluate the similarity of the accessions on microsporogenesis, hierarchical cluster analysis (with squared Euclidean distance and Ward’s agglomeration method) as well as K-means clustering were performed using the length data of each developmental stage of microsporogenesis.

The endodormancy release dates determined by the three methods (microsporogenesis, pistil length measurements and forcing shoots) were compared using two-way Multivariate Analysis of Variance (MANOVA) with factors ‘year’ and ‘accession’ where the length of string and archesporium stages were used. Normality of the variables was checked with Shapiro–Wilk’s test (p > 0.25).

Pair-wise comparisons were performed by Games-Howell’s post hoc test in the case of pistil method and by Tukey’s post hoc test in other cases in order to differentiate among accessions and find the year effect.

The statistical analyses were performed using the R statistical programme version 2.1 (R. CoreTeam, 2021).

The results showed that flower buds of almond accessions underwent the classical developmental stages of microsporogenesis as described in Materials and Methods (Figures 1–3).

Figure 1.

Figure 2.

Figure 3.

In 2019/20, the development of the archesporial tissue ranged between 27 and 64 days from the 1st of November (Figure 1). The string stage, which was marked as the beginning of microsporogenesis process lasted 18–23 days. This stage indicates the transition from the endodormancy to the ecodormancy phase. The transition periods of the PMC and tetrads were short; the anthers remained in the PMC stage from 6 days to 10 days and likewise in the tetrads from 8 days to 11 days depending on the almond accessions. The microspore stage began around 30–45 days after the start of the microsporogenesis process. The transition of the microspore stage lasted between 37 days and 44 days. The end of microsporogenesis (ecodormancy) was between March 9 and April 5. These indicated clearly that the process of microsporogenesis began around 90–100 days before flowering depending on accessions. The accessions differed in the developmental rate of microsporogenesis in particular in showing an important variation in the amount of time taken for the archesporium stage to differentiate into the string stage of microsporogenesis. At later stages of microsporogenesis, the transition periods became shorter and the variation increased during the whole process of microsporogenesis.

In 2020/21, the period of the development of archesporial tissue to produce string cells was shorter compared to the first year, in particular, for the earliest types as the development of the archesporium stage was noted after 15 days from the establishment of dormancy (Figure 2). However, for the case of the latest two cultivars ‘Vairo’ and ‘Constanti’, it remained almost the same. The string stage lasted 9–36 days, while the transition periods of PMC to tetrads and then tetrads to microspores lasted 6–13 days and 8–9 days, respectively. This means that the microspore stage began around 20–60 days after the start of the microsporogenesis process. The period of the development of microspore cells to produce pollen was relatively longer for most accessions compared to the previous season. This stage lasted between 29 days and 83 days. Consequently, the end of microsporogenesis was extended by 5–22 days as the start to blooming date was between March 22 and April 14.

In 2021/22, the speed of microsporogenesis was comparable to that of 2020/21. The development of the archesporial tissue ranged between 14 days and 65 days during this year. However, the transition periods of the string stage were quite short for all the accessions as they lasted between 5 days and 8 days only (Figure 3). The transition periods of PMC and from tetrads to microspores lasted 4–8 days and 8–9 days, respectively. This explains that the microspore stage began around 17–24 days after the start of the microsporogenesis process. But, the development of microspores cells to produce pollen was much slower in this year compared to both the other years as it lasted between 50 days and 90 days depending on accessions. The pollen cells were noticed 19–20 days before the end of microsporogenesis as blooming started between March 21 and April 9. In this year, the process of microsporogenesis began about 90–130 days before flowering similar to that in 2020/21.

As a summary, among all the studied accessions, ‘1/7’, ‘Eriane’ and ‘5/15’ had the shortest period of archesporial in the three studied years. The development of archesporial tissues of these accessions took 27, 15 and 14 days and the development of the microspore stage through tetrads took 71, 39 and 31 days in all three during 2020, 2021 and 2022, respectively. The accession ‘1/7’ came to the end of microsporogenesis process on the 9^th of April, 2 days earlier than ‘Eriane’ and ‘5/15’ in 2020. However, in 2021 and 2022, all the three accessions had their end of microsporogenesis (ecodormancy) on the same day, i.e. on the 22^nd or the 21^st of March, respectively. Accessions ‘7/21’, ‘Constanti’ and ‘Vairo’ had the longest archesporium stage (64, 66 and 65 days in all) during 2020, 2021 and 2022, respectively. The development of PMC to produce the microspores through tetrads took 96, 116 and 89 days in that order. In the flower buds, the final form of pollen grain was noticed 19–24 days before blooming; the blooming started on the 5^th, 14^th and 9^th of April in 2020, 2021 and 2022, respectively.

According to the statistical results, the accessions showed significant differences in their total length and in each developmental stage of microsporogenesis (p < 0.001). Accordingly, accessions were classified based on developmental rates of microsporogenesis and the dendrogram generated by hierarchical cluster analysis with the Ward’s method presented in Figure 4, where 5 main considerable groups and 10 subgroups can be observed for each year. In Figure 4, the green group contains the earliest accessions while the purple consists of the latest ones. The red group can be called ‘early accessions’ while the orange and navy groups are of the ‘medium’ and ‘late accessions’ groups, respectively. We can see that ‘1/7’, ‘Eriane’ and ‘5/15’ were grouped into the same ‘earliest’ group in all the 3 years while ‘35/29 Sóskút’ and ‘Érdi édes’ were classified into the earliest group in 2020 and 2021, but into the ‘early’ group in 2022, which shows the potential climate sensitivity of these two accessions. Note that the K-means clustering with four groups revealed almost the same clustering output as the introduced hierarchical one with the only difference that ‘35/29 Sóskút’ and ‘Érdi édes’ were classified into the ‘earliest’ group in 2022 too. ‘Korai keményhéjú’, ‘Akali 57/2’, ‘Sóskút 96/5’, Tétényi kedvenc’, ‘Sóskút 66/3’, ‘Budatétényi 70’, ‘Tétényi bőtermő’ and ‘Tétényi rekord’ were grouped as early accessions in all the 3 years.

Figure 4.

‘Belona’, ‘Sóskút 16/7’, ‘26/43’and ‘Diósárki’ belong to the medium accessions group while ‘Sóskút 96/1’, ‘Tétényi keményhéjú’ and ‘Marinada’ are the late accessions. ‘Soleta’ was classified as medium accession in 2021 and 2022, while ‘latest’ in 2020. In this difference, K-means and hierarchical clustering agreed which refers to the climate sensitivity of accession ‘Soleta’. Together with accessions ‘7/21’, ‘Constanti’ and ‘Vairo’, ‘6/10’ was also classified in the ‘latest’ group in all the 3 years by both classification methods.

Before the plants entered the endodormancy phase, the growth was continuous and consistent (data not shown). Growth was arrested in the endodormancy phase; the increment was not apparent, with an average length of 1 mm (Figure 5). Pistil growth resumed when archesporium tissues in the anthers of the flower buds differentiated into the string stage, which was considered as a transition from endodormancy to the ecodormancy phase, first at a very slow increment followed by a few days of highly concentrated growth prior to blooming. This indicated that the accumulation of a certain amount of cold required by a cultivar is a prerequisite for pistil growth resumption. The accessions had considerable variations in their pistil growth rate particularly between the early and late types. The growth increment rate after resumption was more rapid of the latest ones than those of the earliest.

Figure 5.

In 2019/20, the almonds presented pronounced variation in their pistil growth rate from the end of November onwards (Figure 5A). On the 27th of November for ‘Eriane’ and ‘5/15,’ after a time period of steady 1 mm length, the rate of growth appeared to increase dramatically to 1.4 mm, after the necessary amount of chill had been accumulated. Also, the pistil length of ‘1/7’ increased slightly at this time although the visible increment was recorded on the 7th of November. This change in pistil growth characteristics was seen after the 16th of January for the latest ones such as ‘7/21’, ‘Vairo’ and ‘Constanti’.

In the years 2020/21 and 2021/22, the resumption of pistil growth of those earliest ones occurred on the 16^th and 24^th of November, respectively. In the case of the latest ones, pistil growth was arrested until the 19th of January in 2020/21 (Figure 5B), while in 2021/22, their growth was arrested further up to the beginning of February (Figure 5C). By that time, the earliest accessions were already with pistil lengths between 4.5 mm and 5.5 mm. In that pistil growth stage, the microspores were distinguishable in the anthers of those earliest ones. Of those earliest accessions, the resumption of pistil growth appeared at the moment when in the anthers of the flower buds, full string stage was distinguishable. While of those latest ones, the pistil appeared to resume growth at the moment when the separate PMC and tetrads in the anthers of the flower buds were distinguishable.

The developmental rates of bud shoots exposed to forcing conditions were clearly affected by accessions and yearly climatic conditions. The flower bud development overlapped into 10 groups with slight differences of overlapping each year (Figure 6). In the year 2019/20, flower buds of the earliest accessions ‘Eriane’, ‘1/7’ and ‘5/15’ started to present open flowers in the first week of November (Figure 6A). The developmental rates were gradual. On the 17th of November, about 22% of the buds presented open flowers. Twenty days later, on the 7th of December, the amount reached 50% and 100% on the 16th of January. By that time, they reached 50% of open flowers and in the anthers of the flowers, 90% of the string was distinguishable. Flower buds opened 100% following the appearance of the microspore stage and pistil growth was already resumed.

Figure 6.

In the case of the latest accessions ‘7/21’, ‘Vairo’ and ‘Constanti’, flower bud opening started after the 6th of January. In the anthers of the flower buds, separate PMC of microsporogenesis stages were distinguishable at this moment. Once started, flower bud opening increased at higher rates as indicated by steep slopes of the sigmoid lines. Of those latest ones, the proportion of opened flowers reached half in the first week of February and 100% towards the end of the month. This was 30 days before flowering under field conditions. During this year, the flower buds of forced shoots for most of the studied accessions presented around 50% open flowers with the appearance of tetrads and microspore stage.

In the year 2020/21, flower buds of the earliest accessions began to present open flowers after the 1st of December, and 15 days later, about 50% of the buds presented open flowers (Figure 6B). The flower bud opening began with the appearance of tetrads and in the anthers of the flower buds. The end of the reduction process was already indicated by the occurrence of the microspore stage at the time when about 50% of the buds presented open flowers. For the case of the latest ones, ‘7/21’, ‘Vairo’ and ‘Constanti’, flower bud opening began after the 6th of January with the string stage formation. Around 50 days later, on the 26^th of February, the proportion of opened flowers reached more than half when in the anthers of the flower buds the microspores and the first change of pistil length increment rates were noticeable. Flower buds of forced shoots began to present open flowers with the appearance of tetrads and presented around 50% open flowers with the appearance of a microspore stage similar to that of 2019/20. However, with some accessions, flower bud opening started with the transition to the string stage and presented more than half-open flowers with the development of tetrads.

In the year 2021/22, the development of flower buds under the forcing started a bit earlier than in 2020/21 but similar to that of 2019/20. The rate of flower bud opening was quicker in this year than in the years 2019/2020 and 2020/21 (Figure 6C). On the 15^th of November, about 25% of the buds of the earliest accessions ‘Eriane’, ‘1/7’ and ‘5/15’ presented open flowers when the archesporium tissue in the anthers of flowers differentiated to the string stage. Seven days later, the flower buds presented 50% open flowers with the formation of tetrads. By that date, the flower buds of these accessions presented around 38% and 1% of open flowers in 2020 and 2021, respectively. This is probably due to differences in environmental factors observed in the 3 years that influenced the response of buds to forcing temperatures and flowering date in spring. For most of the studied accessions, the flower bud opening started with the formation of the string stage and reached around 50% open with the appearance of separate PMC and tetrads during this year. This is probably due to differences in environmental factors, especially temperature.

After all, the dates of endodormancy determined by microsporogenesis, pistil length measurements and forcing shoots were compared using two-way MANOVA with the factors ‘year’ and ‘accession’.

The overall MANOVA resulted in significant accession and year effect (Wilk’s lambda = 0.006 and 0.023, respectively, both with p < 0.001).

The follow-up univariate Analysis of Variance (ANOVA) revealed highly significant genotype and year effect for all the three methods (accession: F [24; 48] > 28.8; year: F [2; 48] > 13.71, all with p < 0.001). Post hoc test results show that all three methods were able to differentiate the accessions, however, showed different resolutions and slightly different classifications of the accessions (Table 1). Most groups (eight) was found among microsporogenesis studies, whereas three groups were formed by pistil length and two forcing method. They all agree that accessions ‘1/7’, ‘Eriane’ and ‘5/15’ are significantly earlier than all the others. However, based on the pistil method, accessions ‘35/29 Sóskút’, ‘Érdi édes’, ‘Akali 57/2’and ‘Tétényi bőtermő’ also appear among the earliest. Similarly, according to pistil length measurements, the accessions ‘6/10’, ‘Soleta’, ‘7/21’, ‘Constanti’ and ‘Vairo’ are significantly later. But based on the other two methods, accessions ‘7/21’, ‘Constanti’ and ‘Vairo’ are the significantly later ones.

As for the year comparisons, post hoc tests agreed that the year 2021/22 resulted in the earliest development based on all the three methods. However, while in the case of microsporogenesis method, the development was significantly earlier in this year than in 2020/21, and in the year 2019/2020, it was significantly later than both of the other years; the order was different in the case of forcing method: the significantly latest year was 2020/21. In the case of pistil method, years 2021/22 and 2020/21 did not differ significantly, while 2019/2020 resulted in significantly later development.