JARS v37n2 - The Timing of Microsporogenesis in Deciduous Azalea

The Timing of Microsporogenesis in Deciduous Azalea ¹
Mark P. Widrlechner, Harold M. Pellett, and Peter D. Ascher
Research Assistant and Professors, respectively
Department of Horticultural Science & Landscape Architecture
University of Minnesota, St. Paul, MN

Editors note:
The research paper below is a report of research supported by funds from the Research Foundation, the research arm of the American Rhododendron Society. It deals with the problem of making chromosome counts in deciduous azaleas. Almost no research has been done in this area in the last 20-30 years because of the difficulty of collecting the bud material at the proper time. The research attempted to improve on present methods by studying the effects of temperature. Although the heat units method studied here improved the estimation of proper timing of collecting bud material, the authors suggest further research - that in addition to, time of year and accumulated effects of temperature, it is necessary to also consider water stress. Chromosome counts made from bud tissue are preferable to counts made from other parts of the plant because they are more accurate and reveal information on the breeding behavior and genetic background of the plant.

Chromosome counts are desirable also to verify some earlier reports which are now coming under closer scrutiny. For example Rhododendron occidentale was first reported to be a diploid with 26 chromosomes. There is a distinct possibility that the identity of the plant from which the earlier counts were taken was incorrect, because a later report showed R. occidentale to be a hexaploid with 78 chromosomes. (See April 1975 ARS Quarterly Bulletin for a discussion of this matter.)

For interest of members of the Society, Mark Widrlechner has just now completed requirements for his doctorate degree.

Introduction
The study of meiosis in deciduous azaleas is hampered by a lack of published information on the timing of sporogenesis. Because of morphology, microsporogenesis is easier to observe than megasporogenesis and the few reports describing meiosis in the genus Rhododendron L. (Li, 1957; Sax, 1930) have studied microspore formation only.

Experiments performed under controlled conditions on both evergreen and deciduous azaleas (Naskali, 1961; Skinner, 1939) suggest that flower development is dependent on temperature and day length. But two reports on the timing of microsporogenesis under field conditions (Bowers, 1930; Goto, 1968) give no indication of the interaction between environmental conditions and microsporogenesis. Therefore, our study was undertaken to examine the influence of day length, estimated heat units, and number of days post-flowering on the timing of microsporogenesis in deciduous azaleas.

Materials and Methods
To determine the timing of microsporogenesis, plants grown outdoors at the University of Minnesota Landscape Arboretum, Chanhassen, MN, were examined at two to four day intervals during August and September 1981 and 1982. Table 1 lists the plants examined each year. For these plants we also recorded the approximate dates of peak flowering and microsporogenesis.

Table 1. Rhododendron Taxa Sampled to Determine the Timing of Microsporogenesis
Taxon	Minn. Landscape Arb. Accession Number	Years Sampled
R. calendulaceum (Michx.) Torr.	570085 B	1981 & 1982
R. calendulaceum O.P.	800064	1981 & 1982
R. calendulaceum x 'Exbury Hybrid’	800096	1981 & 1982
R. canadense (L.) Torr.	621077	1981 & 1982
R. x kosterianum Schneider x R. prinophyllum (Small) Millais	570494 K	1981 & 1982
( R. x kosterianum x prinophyllum ) x ( R. x kosterianum x prinophyllum )	690331	1981 & 1982
R. canadense x R. × kosterianum	620014	1981 only
R. x kosterianum x R. prinophyllum	630002 Q	1981 only
R. prinophyllum x R. x kosterianum	671039C	1981 only
R. atlanticum Rehd.	570091 A	1982 only
R. prinophyllum x R. x kosterianum	671039 B	1982 only
R. viscosum (L.) Torr.	790153	1982 only

Inflorescence buds were collected in the morning and immediately fixed in a modified Carnoy's Solution [Absolute Ethanol: Glacial Acetic Acid: Chloroform (3:1:4)] (Brooks et al., 1950). On each collection date three to five buds were sampled, except in cases where the plants did not produce enough buds. After the buds were fixed, we removed individual flower primordia from the inflorescence buds and dissected out at least ten anthers from the primordia. Anther length was measured using an ocular micrometer in a compound light microscope and, if the anthers appeared to be sufficiently developed, they were then squashed and stained with aceto-carmine (Brooks et al., 1950), to determine whether microsporogenesis had occurred.

Data on daily maximum and minimum temperatures for the summers of 1981 and 1982 were obtained from records maintained at the nearby University of Minnesota Horticultural Research Center. These data were converted to heat units using four variations of the remainder system (Table 2) (Shaw, 1977). Methods A and C are uncorrected for extreme temperatures. Method A uses a base temperature of 43°F (6°C), following Jennings' (1979) study of development in Rubus and method C uses a base temperature of 50°F (10°C), as recommended in studies of Zea (Shaw, 1977). Methods B and D are modifications of methods A and C, respectively, which set all daily minimum temperatures below a given base temperature at that base temperature, and which compensate for the possible deleterious effects of temperatures above 86°F (30°C) by reducing extreme daily maximum temperatures, modifications used by Cross and Zuber (1972).

Table 2. Methods Used to Compute Heat Units

Method A: (0.5 x (DT max + DT min ))-43°F Method B: If DT max < 86°F and DT min > 43°F, then (0.5 x (DT max + DT min ))-43°F If DT max < 86°F and DT min > 43°F, then (0.5 x DT max )-21.5°F If DT max >86°F and DT min > 43°F, then (0.5 x (DT min -DT max ))+43°F If DT max >86°F and DT min <43°F, then 64.5°F -(0.5 x DT max ) Method C: (0. 5x (DT max + DT min ))-50°F Method D: If DT max < 86°F and DT min > 50°F, then (0.5 x (DT max + DT min ))-50°F If DT max < 86°F and DT min <50°F, then(0.5 x DT max )-25°F If DT max >86°F and DT min > 50°F, then (0.5 x (DT min - DT max ))+ 36°F If DT max >86°F and DT min <50°F, then 61°F - (0.5 x DT max ) DT max = Daily Maximum Temperature DT min = Daily Minimum Temperature Note that for all methods, only positive heat units are accumulated.

The accumulation of heat units was calculated from two starting dates. The first method begins with the date of peak flowering, an easily observable characteristic and also one of biological significance because it is at this point that fruit production begins. The other method, which begins with the summer solstice, was chosen primarily because decreasing day length causes important physiological changes in woody plants (Wareing and Phillips, 1970) and because Batta (1974) found August temperatures to be positively correlated with flower production the following spring. Starting heat unit accumulation on 21 June generally increased the proportion of August heat units in the total.

Mean deviations for these methods of determining the timing of microsporogenesis were calculated by averaging the 1981 and 1982 tests for four genotypes measured in both years. The dates on which the average test value occurred for that genotype were compared with the actual dates on which microsporogenesis took place. In future tests, these deviations may be used to determine the potential predictive value of the methods.

For example, using method A, for R. x kosterianum x prinophyllum accession number 570494 K, meiosis took place on 8 August 1981, 1951.5 heat units after peak flowering. In 1982, it took place on 21 August, 2217.5 heat units after peak flowering. The average number of method A heat units was 2084.5. In 1981, 2084.5 heat units were accumulated by 13 August and in 1982 this total was reached on 17 August. The average value of method A heat units, measured after peak flowering, was reached within four or five days of meiosis.

Results and Discussion
Data collected in 1981 and 1982 on the timing of microsporogenesis are presented in Table 3. The timing varied greatly among and within the taxa examined. The earliest occurrence was for R. canadense in 1981, when tetrads were already present on 6 August, the first collection date. The latest was also in 1981, on 21 September, for R. calendulaceum OP. A frost occurred seven days later. These data do not conflict with Bowers' (1930) findings that microsporogenesis takes place about the first week in September for R. catawbiense Michx. in New York, nor do they deviate from Goto's (1968) table showing a range of 20 July to 20 October for microsporogenesis in 18 Rhododendron species in Japan, with all but three forming pollen between 6 August and 21 September.

Average anther size at microsporogenesis is given in Table 4. In the four cases where two-year comparisons can be made, there is little variation within genotypes. Once the characteristics of a genotype are known, anther size may then be used in subsequent seasons to indicate when to collect buds to observe meiosis.

Based on limited data (two years and four genotypes), Table 5 lists the mean deviation in days between the observed date of microsporogenesis and the calculated date, using the various methods of estimating the timing of microsporogenesis. The smaller the deviation, the better may be a method's potential predictive value. The mean deviation when microsporogenesis was predicted using the number of days after 21 June is 5.4 days. This method assumes that microsporogenesis is primarily controlled by day length. The deviation when peak flowering was used as a beginning date is 4.3 days. This method assumes that there is a relationship between developmental stage and microsporogenesis independent of normal environmental fluctuations at the test site.

Heat unit methods account only for temperature. In contrast to Batta's (1974) results, limiting the accumulation of heat units to after 21 June did not decrease the mean deviation. Average heat unit deviation after peak flowering equals 4.3 days, but after 21 June it equals 4.9 days. The compensation for extreme temperatures used in methods B and D may have had some effect. Mean deviation with compensation equals 4.2 days, which is 0.9 days better than that obtained without compensation. The base temperature also appeared to have some influence, with 43°F (mean deviation = 4.2 days) being more effective than 50°F (mean deviation = 5.1 days). Of all methods, heat unit method B, which used a base temperature of 43°F and extreme temperature compensation, with accumulation beginning after the date of peak flowering, had the smallest deviations. However, the mean deviation was only one day less than that produced by the days after peak flowering method, a much simpler method. In future studies, heat unit accumulation methods may be improved by testing base temperatures and extreme temperature corrections that correspond better to the metabolism of azaleas, rather than relying on methods developed for other plants.

Another possible way to improve on the methods we have used was suggested by Wielgolaski (1974), who noted that "for most species, however, both herbaceous and woody, increasing precipitation was more or less favorable to rapid plant development again in the latest periods, i.e. from flowering onward." The addition of some measure of water stress to heat unit methods, such as Caprio's (1971) solar-thermal unit method, might be more accurate. In support of some modification, we note that from 1 June to 21 August of 1982 there were only 4.36" (111 mm) of precipitation at the Horticultural Research Center, less than 40% of the precipitation of the same period in 1981. Plants, such as R. canadense and R. x kosterianum x prinophyllum , that develop buds early in the summer, were later in 1982, both in calendar days and in heat units.

Acknowledgements
We sincerely thank Doctors Charles Burnham, David Davis, James Luby, and Carl Mohn for their helpful suggestions, Brent Pemberton and Shosuke Kaku for pointing out references, Dave Bedford for providing weather data and Sue Fuhrman for helping collect azalea buds. The financial assistance of the American Rhododendron Society Research Committee is greatly appreciated.

Literature Cited
Batta, J. 1974. Blühverlauf und win-terschäden bei rhododendron. Rhod. lmmergr. Laubg. Jahrb. 1974: 58-67.
Bowers, C.G. 1930. The development of pollen and viscin strands in Rhododendron catawbiense . Bull. Torrey Bot. Club 57: 285-314.
Brooks, R.M., M.V. Bradley, and T.I. Anderson. 1950. Plant Microtechnique Manual. Davis, CA: Univ. of California.
Caprio, J.M. 1971. The solar-thermal unit theory in relation to plant development and potential evapo-transpiration. Mont. St. Univ. Agric. Expt. Sta. Circ. 251.
Cross, H.Z. and M.S. Zuber. 1972. Prediction of flowering dates in maize based on different methods of estimating thermal units. Ag-ron. J. 64: 351-355.
Goto, T. 1968. Tsutsuji Satsuki. Tokyo: Kashima-shoten.
Jennings, D.L. 1979. Genotype-environment relationships for ripening time in blackberries and prospects for breeding an early ripening cultivar for Scotland. Euphytica 28: 747-750.
Li, H.L. 1957. Chromosome studies in the azaleas of eastern North America. Amer. J. Bot. 44: 8-14.
Naskali, R.J. 1961. Initiation and early ontogeny of the inflorescence of Rhododendron obtusum . M.S. Thesis, Ohio St. Univ.
Sax, K. 1930. Chromosome stability in the genus Rhododendron . Amer. J. Bot. 17: 247-251.
Shaw, R.H. 1977. Climatic requirement. In. Corn and Corn Improvement, 2nd ed. G.F. Sprague, ed. Madison, WI: Amer. Soc. Agron., pp. 591-623.
Skinner, H.T. 1939. Factors affecting shoot growth and flower bud formation in rhododendrons and azaleas. Proc. Amer. Soc. Hort. Sci. 37: 1007-1011.
Wielgolaski, F. 1974. Phenology in agriculture. In. Phenology and Seasonality Modeling. H. Leith, ed. New York: Springer-Verlag, pp. 369-381.

¹ Scientific Journal Series Paper No. 13,222, Minnesota Agricultural Experiment Station.