American Journal of Botany, 43(1): 7-12 (Jan., 1956)
G. J. Von Abrams and M. E. Hand

1 Received for publication April 18, 1955.
The statistical treatment of this material has been prepared independently by Dr. L. D. Calvin, Oregon State College. This cooperation is gratefully acknowledged.

THE SEEDS of many plants are known to be dormant at maturity, and to require a period of exposure to low temperature to enable them to germinate under subsequent suitable conditions. In Rosa, and in several other rosaceous genera, this low-temperature after-ripening is widely regarded as a fixed and characteristic prerequisite to further development of the embryonic plant. Joseph (1929) pointed out that rosaceous seeds require low temperature stratification to prepare them for germination. Haut (1938) found that apple, cherry, peach and pear seeds must be after-ripened before germination will occur. Crocker (1927) reported that seeds of temperate-zone roses germinate well only following a period at low temperature, and Crocker and Barton (1931) determined the after-ripening requirements of hybrid and species rose seeds, as well as those of several other rosaceous seeds.

The ubiquitous nature of the low-temperature after-ripening requirement of rosaceous seeds is apparently unquestioned in recent literature, with the exception of a report by Calvino (1930) stating that hybrid rose seeds grown at San Remo, Italy, germinated readily with no prior low-temperature treatment. This report was accorded little significance, and the results were logically explained by Crocker and Barton (1931) by supposing that the seed were inadvertently exposed to winter temperatures sufficiently low and prolonged to allow after. ripening to occur. More recently, however, there have been unpublished reports from southern France, and from some of the inland valleys of California, claiming successful germination of hybrid rose seed without resort to low-temperature treatment. These reports have been partially substantiated in this laboratory by the collection or import of seed from several localities, and in the course of the investigation it has become apparent that the dormancy of hybrid rose seed of a given genetic origin may vary widely from year to year, and from one region to another. The present paper deals with the quantitative aspect of the dormancy, and with the relationship between the dormancy and the pre-harvest climatic environment.

MATERIALS AND METHODS.—Temperature data were obtained in the experimental field from shielded maximum and minimum registering thermometers. Sunlight data were calculated from the Local Climatological Data of the U. S. Weather Bureau station at Portland, Oregon. This station is located approximately 20 mi. from the experimental area. The climatic data presented in table 1 represent two periods of time: the 30-day period and the 60-day period immediately preceding the date of harvest, November 1. The average daily temperature was calculated as the average of the daily maximum and minimum Similarly, the average 8-hr. day temperature was calculated as the average of the maximum and minimum reached between the hours of 8:00 AM. and 4:00 P.M., Pacific Standard time, and the average 16-hr. night temperature was calculated as the average of the maximum and minimum reached between 4:00 P.M. and 8:00 A.M. The latter period was subject to natural daylight and was not a continuous dark period. Hourly temperature records are not available.

The controlled hand pollinations prerequisite to these studies, involving ten representative crosses between 16 varieties (table 2), were made during the last seven days of June of each year. The parent varieties used may be identified to the following extent. Numbers 125, Pinocchio, and 110, Else Poulsen, are assigned to the horticultural Floribunda class. Numbers 39, 24, and 21 are, respectively, hybrids of R. eglanteria, R. wichuraiana and R. rugosa with hybrid Teas. Number 12, Crimson Glory, number 38, Mirandy, and the remaining varieties are hybrid Teas of varied complexity. It may be noted that there is a reciprocal cross represented, and that one seed parent participates in two different crosses, as does one pollen parent.

The fruit were harvested on the first day of November. Seeds were immediately removed from the fruit, cleaned of extraneous tissue, and washed. They were held one minute in a dilute detergent solution containing 0.5 per cent NaOC1. Those seeds which did not sink readily in this solution were either discarded or handled separately, since this behavior generally indicates an aborted or degenerating embryo. The seeds were planted on one-inch squares in flats containing a sterile medium composed of equal parts of sand, finely ground peat and commercial expanded aluminum silicate. This medium was kept moist throughout the study. The seed-flats were held for 60 days in a greenhouse at a minimum 8-hr. day temperature of 21°C., and a minimum night temperature of 15°C. These conditions are suitable for the germination of the nondormant seeds. The initial germination period was followed by 90 days in a cold-cabinet at 2° ± 0.50 C., a second 60-day germinating period in the greenhouse, a second 90-day cold period, and a final 60-day germinating period. The duration of the various periods is not considered to be necessarily optimum. Any observed rupture of the seed-coats followed by elongation of the plant axis is credited in these data as a successful germination, without con­sideration of the subsequent normality or viability of the seedling.

It is necessary to acknowledge a number of variables, other than pre-harvest climate, which might be expected to exert an influence upon the results obtained. Some variation in germination is found from one fruit to another of the same cross, even from the same plant. Accordingly, the number of fruit representing each sample is greater than 100. There is occasionally a significant difference in germination between seed crops of the same cross obtained from different individual seed-bearing plants. In order to minimize the influence of this variability, 15 seed-bearing plants were used for each cross, unless otherwise noted. The number of fruit supported by the individual plants within a variety was kept approximately equal from year to year during the entire period of the study, in anticipation of a competitive relationship. Within the time limits of this work, it appears that age of the seed-bearing plant has little effect upon germination. It is difficult, however, to separate effectively differences in age from circumstantial differences in general condition, and in several cases there have occurred significant differences in germination which can not be proven independent of this variable. Consequently, the results reported for each year are those obtained from new plantings of seedbearing plants. These plants were propagated in the same area and in the same manner, bud-grafted on rooted cuttings of a strain of R. multiflora. They were selected according to size and conformation at two years of age, and planted in one uniform block in March of the year in which they were used for the present study. Variation in the prevalence and intensity of attack by fungi and insects was well controlled by regular application of ferric dimethyldithiocarbamate, elemental sulfur and benzene hexachloride. Satisfactory control of nutrition in the field was not possible. Nitrogen, phosphorous and potassium were maintained at moderate levels, as determined by periodic soil analysis. Parallel experiments with high nitrogen or high phosphorous during August and September of 1950, and again in 1951, did not result in increased germination. Soil moisture was kept at a high level during the entire investigation.

2 This terminology would appear to have been established by
precedent for rosaceous seeds. The cotyledons are included.

In order to measure the potential germinability of the embryos, 100 seeds of a given cross were withdrawn randomly from those which had been prepared for planting. The stony pericarps and the integumentary tissues were removed, and the excised embryos2 were cultured aseptically in vitro on a non-nutrient medium. Acid-washed hard glass vessels were used, and the medium was moistened with glass-redistilled water. Successful germination was conceded, for this purpose, when the cotyledons had spread to an angle of 45° or more from the axis, and the primary root had attained a minimum length of 5 mm. The maximum time allowed for germination was 14 days.

As a preliminary examination of the effect of controlled preharvest climate upon germination, six plants each of varieties number 12, 21 and 30 were planted in 20-in, perforated waxed wooden tubs on March 14, 1953. The tubs were inserted to ground level in the experimental field. These plants were handled in the manner previously described until the first day of October, when the tubs were removed from the ground. Three plants of each variety were then replaced in their former positions, and the others were moved to a greenhouse cubicle in which a minimum 8-hr. day temperature of 21°C. was provided. A day temperature of 23°C. was exceeded upon three occasions, when the temperature rose to maxima of 28°, 30° and 28°C., respectively, on 4, 5 and 7 October. Night temperatures were held at a minimum of 15°C. Batteries of white fluorescent tubes were used to maintain a minimum of approximately 180 ft.-c. near the tops of the plants during the 8-hr. high-temperature period. Soil moisture was kept high. The fruit of both greenhouse and control plants were harvested on the first day of November, and the seeds were handled and planted in conformity with the method previously described. For each of the three crosses, 225 seeds were selected randomly from the greenhouse plants and the same number from the control plants. Once harvested, all these seeds were kept under greenhouse conditions, and were not exposed to temperatures below 15°C.

RESULTS.—Climate.—The observations presented in this report were made possible by a circumstantial variation in local autumnal climate over the 5-yr. period, 1949 through 1953. Narratively, the months of September and particularly October, a period in which physiological maturation of the seed must occur in this area, were characterized by progressively higher temperatures from 1949 through 1952. A decline occurred in 1953, to a level between that of 1951 and 1952. The points of temperature reached in this progression represent significant increments, as may be seen in table 1.

Germination.—Table 2 shows the successful germination achieved by 900 seeds of each of ten crosses for each year of the study. In each case, the 900 seeds were selected randomly from a larger sample. The upper line of data for each cross in table 2 indicates the germination which occurred during the 60-day period prior to any exposure to low temperature; the next lower line indicates the total germination up to the end of the second 60-day germinating period, which followed 90 days of exposure to a temperature of 2°C.; the third line indicates the total germination after three 60-day germinating periods and two 90-day low-temperature periods, a total elapsed time after planting of one year. The correlation coefficient relating germination to the mean of the average daily temperatures for the 30 days preceding harvest has been included for the total germination of each cross.

TABLE 1. Climatic data for pre-harvest period

Period Factor 1949 1950 1951 1952 1953
30 days Average daily max. T° C 14.2 15.7 16.8 22.4 18.7
      min. T° C 4.3 5.8 7.6 7.5 7.1
      range, T° C 9.9 9.9 9.2 14.9 11.6
  Mean of average daily T° C 9.2 10.7 12.2 14.9 12.9
      8-hr. day T° Ca 11.7 13.9 14.7 19.0 15.3
      16-hr. night T° C 8.6 9.6 13.6 18.3 11.2
  Sunshine, total hoursb 106.8 43.6 116.9 201.7 137.3
60 days Mean of average daily T° C 13.2 13.6 14.9 16.6 15.2
      8-hr. day T Ca 16.1 17.4 18.2 21.0 17.9
  Sunshine, total hoursb 311.4 290.1 357.1 474.9 342.8

a 8:00 AM, to 4:00 P.M., Pacific Standard time.
b Calculated from Weather Bureau records, Portland city station, Oregon.

It is apparent that the germination of each of the ten crosses increased progressively from 1949 through 1952, followed by a regression for the crop of 1953. Germination without low temperature after-ripening occurred in all crosses in the 1952 crop, but did not occur at all in the crops of 1949 and 1950. Results during the first two germinating periods show no deviation from the pattern established by the total germination.

  TABLE 2. Per cent germination of intact seeds. N = 900

Cross Low T°
1949 1950 1951 1952 1953 ra
30 125 0 0 0 6.2 42.3 13.2  
90 4.2 12.7 26.9 60.1 35.2  
180 12.8 27.7 44.3 66.9 49.4 0.998
110 35 0 0 0 4.7 33.9 8.4  
90 5.2 15.4 26.1 43.0 30.0  
180 15.8 25.4 37.0 51.6 40.9 0.996
38 19 0 0b 0 0 14.7 0  
90 2.1b 5.1 12.1 30.6 15.8  
180 7.9b 18.3 29.1 45.3 34.8 0.996
24 46 0 0 0 2.7 21.0 4.2  
90 2.8 6.1 14.1 29.6 16.8  
180 7.2 14.9 24.6 38.2 27.3 0.999
12 26 0 0 0 0 14.2 0.8  
90 0.6 6.0 12.8 25.2 16.1  
180 2.4 11.2 20.9 35.6 23.4 0.999
12 23 0 0 0 0 12.6 0.2  
90 0.2 4.7 10.7 18.9 12.3  
180 0.6 7.8 17.0 30.4 20.1 0.999
21 11 0 0 0 0.7 17.7 2.1  
90 0.9 7.2 15.6 30.1 18.8  
180 4.8 12.8 19.3 33.8 24.4 0.998
46 24 0 0 0 0 7.9 0  
90 0 0.2 4.3 14.2 8.1  
180 0 1.3 8.2 20.1 11.4 0.999
39 15 0 0 0 0b 4.9 0  
90 0.7 4.3 6.9b 15.9 9.2  
180 1.1 7.6 15.0b 27.1 18.6 0.999
20 24 0 0 0 0 1.3 0  
90 0.1 1.7 3.0 6.3 3.8  
180 1.2 4.4 9.2 15.2 9.9 0.997

a Correlation coefficient with the mean of the average daily temperatures for the 30-day pre-harvest period.
b Results from only 13 seed-bearing plants.

The relationship between germination and preharvest temperature is shown graphically in fig. 1. The total germination for each of three crosses is plotted against the 30-day mean of the average daily temperatures, since this factor provides the best correlation of any of the available climatic data. An excellent straight-line relationship is indicated over the entirety of the temperature range. If the same results are plotted against the 30-day mean of the average 8-hr. day temperature, the curves are found to be sigmoid.

Excised embryos.—The germination of excised embryonic plants in vitro is shown, for four crosses, in table 3. The results obtained from any one of the crosses were remarkably uniform from one year to another. No difference was observed between the crops of 1951, 1952 and 1953 in the time required for germination, nor in the vigor of the young plants. Embryos from the 1949 and 1950 crops, however, germinated somewhat more slowly.

Supplemental pre-harvest temperature and light.

Fig. 1. The germination of three crosses of hybrid rose seed from crops of five successive years, as related
to the mean of the average daily temperatures prevailing during the 30-day period preceding the harvest.
Curve 1: cross 30 x 125. Curve 2: cross 21 x 11. Curve 3: cross 20 x 24. N = 900.

TABLE 3. Per cent germination of excised embryos in 14 days. N=100

Cross 1949 1950 1951 1952 1953
30 125 97 96 99 98 98
12 26 93 95 96 95 93
21 11 98 96 100 97 99
20 24 81a 89 85 88 86

a 9 embryos contaminated and discarded.

—Plants which were transferred from the field to the greenhouse, where an 8-hr. high-temperature photoperiod was provided during the 30-day period preceding harvest, produced seeds characterized by a relatively slight dormancy as compared to seeds obtained from control plants remaining in the field. The table 4 presents data demonstrating this response for three crosses of the 1953 crop. Germination occurring in the 60-day period after harvest, without low-temperature after-ripening of the seeds, was remarkably greater in those samples obtained from plants which had been subjected to the controlled climate. Table 1 shows that the mean average daily temperature recorded in the field for this 30-day period was 12.9°C., and the mean average 8-hr. day temperature was 15.3°C. Weather Bureau records show an average daily sunshine, for the period, of 4 hr. and 35 min. In contrast, the mean average daily temperature in the controlled cubicle was above 18°C., and the 8-hr. day temperature was held at a minimum of 21°C., with concurrent supplemental light.

DISCUSSION.—It is apparent that the dormancy of hybrid rose seeds, and their consequent germination pattern in time, may vary widely from one year to another within the hereditary design. This is clearly shown by the germination data presented in table 2. In eight of the ten crosses, the range of variation in total germination from 1949 to 1952 was greater than 25 per cent. There is good evidence, not dealt with here, that an equivalent variation may occur between the geographic areas in which seed development takes place. This quantitative aspect of the dormancy is not an unexpected one, but it is easily neglected in the course of physiological studies as well as in efforts to determine a general optimum procedure for after-ripening and for germination.

TABLE 4. Effect of 30 days of supplemental pre-harvest temperature and light, 1953. Per cent germination in 60 days, without low temperature treatment. N = 225

Cross Pre-harvest environment
Field Greenhouse
30 125 14.2 55.6
12 26 0.9 31.1
21 11 1.8 23.1

That the condition is indeed a form of seed dormancy, rather than a degeneration or deficient growth capacity of the embryo, is well demonstrated by the results of those experiments in which excised embryonic plants were successfully germinated in aseptic culture. The medium used was essentially one providing no extrinsic nutrition nor stimulation. The embryonic plant, therefore, may be concluded to be physiologically, as well as morphologically, equipped for the growth process. It is of interest to note, however, that it has not been proven of the intact dormant seed that the failure to germinate is not the result of some physiological delinquency or lassitude of the embryo, not directly concerned with growth per se. For example, there is tentative evidence indicating that it may lie within the functional scope of the embryo of the non-dormant seed to effect a chemical alteration of the suture of the stony pericarp. The high uniform viability of the embryos is surprising in consideration of the poor germination which is usually reported for hybrid rose seed. The variation in the dormancy and germination of intact seeds is not expressed in the germination of the excised embryonic plants. This is of interest in regard to the statement of Crocker and Barton (1953) that Flemion has found the non-after-ripened rosaceous embryo sluggish in its ability to absorb and transport water. They suggest that the transformation of large molecule compounds to more soluble smaller molecule compounds during low-temperature after-ripening results in a greater growth force and provides a more readily available nutrition for the embryo. The embryos used in the present experiment did not appear to suffer any significant deficiency, although those from the most dormant crop were slightly slower to germinate than those from the least dormant crop.

It is obviously desirable, both from the theoretical and the applied points of view, to learn the cause of the variation in seed dormancy, and to be able to reproduce the extremes at will. The evidence is very strong that the dormancy is closely related to the climate prevailing during the period in which the fruit and the seed reach an advanced stage of development. It may be objected that the control obtained in the field experiments was not rigorous. However, in consideration of the general parallelism found between germination and particular climatic factors over the 5-yr. period, including the discontinuity in time represented by the regression of 1953, it seems most unlikely that the relationship is a fortuitous one. The validity of the proposal is further enhanced by the results obtained from the experiment in which seed-bearing plants were transferred to a controlled greenhouse. The results in this case were conclusive, and the possibility of environmental variation other than climate was greatly reduced.

This has been, of necessity, an exploratory study, and the limitations imposed by its methods do not permit identification of the particular climatic factor or factors responsible for the effect upon dormancy. It is apparent that temperature is somehow involved, and this is to be expected. It is also anomalous, however, for after-ripening, which leads to the same end result, is usually considered to be dependent upon low temperature. Of the climatic factors for which some measurement is available, the best correlation is obtained with the mean of the average daily temperatures over a period of 30 days preceding the date of harvest. The excellence of this correlation, as shown in table 2 and graphically for three crosses in fig. 1, is remarkable, and very possibly misleading, for the average of daily extreme temperatures is not physiologically definitive. It has been pointed out that if the total germination is plotted against the mean of the average temperatures of an 8-hr. day (8:00 A.M. to 4:00 P.m.) over the same 30-day period, the curve is sigmoid. A curve of this sort is certainly as plausible as the straight line relationship. If the effect upon dormancy is directly dependent upon photosynthetic activity during a critical period, the 8-hr. average temperature, within the range in which this work was done, should more accurately reflect the relationship than the 24-hr. average. At present it can only be said that tentative evidence does not indicate that extensive photosynthesis is essential, since half-defoliation of the entire plant for the 30 days preceding harvest does not appear to reduce the influence of temperature upon dormancy of these seeds.

Went (1953) has frequently emphasized the importance of night temperature to phasic plant metabolism, and of the relationship between light and dark period temperatures. These climatic factors may well be involved in the effect upon seed dormancy, although the inadequate field data do not show this to be the case. The 30-day average daily minimum temperature, the average daily temperature range, and the mean of the average 16-hr. night temperatures, given in table 1, appear to be unrelated to the germination pattern. This is also true of the only available light measurement, which is expressed in Weather Bureau records as hours and minutes of sunshine. Light intensity data are not available.

Whether or not there is in fact a critical period during which climate is able to influence seed dormancy, following some prerequisite phase of development, or whether the influence is simply an accumulative one, is not elucidated by these data. Went (1953) has justifiably criticized the use of the heat-sum concept in relation to growth, but presumably the process by which dormancy is overcome or by-passed in the hybrid rose seed is not directly concerned with growth. It is apparent from results of the controlled temperature-light experiment that exposure of the seedbearing plants to these conditions for the 30 days preceding harvest was sufficient to reduce seed dormancy significantly.

Tentative evidence indicates that the dormancy may be further reduced, under these same controlled conditions, by postponing harvest and thus extending the total development period. The dormancy may prove to be expressable, within limits, as an inverse function of pre-harvest climate-time. The effectiveness of climate during earlier periods of seed development is not known, but it is apparent that the climatic data for the 60-day period preceding harvest are not as well correlated with the germination pattern as are the data for the 30-day period. There may be a stage of seed or fruit development which must be attained before the dormancy becomes subject to the climatic influence. It is evident that the highest germination obtained from intact seeds is far less than that attained by excised embryos. Whether an optimum pre-harvest climatic environment would produce seeds capable of this high level of germination is not known. There is no indication of temperature satiation in the present data, but it is expected that other factors may become limiting at higher temperatures.

There is little precedent in the literature to support an hypothesis linking the dormancy of a rosaceous seed with pre-harvest climate. One relationship which has been reported with interesting frequency is that between climate and the fat metabolism of various seeds. The results of Painter et al. (1944) indicate that climate affects both the relative quantity of oil in flax seed and the degree of unsaturation. The iodine number was found to be lower in the oil of seed developed under high temperatures. Flemion (1933), working with the rosaceous Rhodotypos kerrioldes, reported a decrease in total seed fats as after-ripening progressed. It is of interest to compare the observations of Harrington and Thompson (1952) to the present work, since they found the area in which lettuce seed were produced to have a strong influence upon germination. Germination at temperatures above 25°C. was concluded to be significantly related to the average temperatures over periods of 10 or 30 days before harvest.

The results of the present study show conclusively that hybrid rose seed do not invariably require low temperature after-ripening for normal germination. Indeed, of the 1952 crop, each of the ten crosses attained significant germination without exposure to low temperature. The results obtained from seed-bearing plants held under controlled greenhouse conditions substantiate this conclusion. It is quite possible, then, that the report by Calvino (1930), claiming germination without after-ripening, was valid. In consideration of the climatic characteristics of southern Italy, where Calvino's work was done, it would seem that the germination obtained might very well have been a result of prolonged high temperatures preceding harvest, rather than a result of inadvertent after-ripening as suggested by Crocker and Barton (1931). Crocker and Barton might have suspected the relationship be­tween seed dormancy and climate, since they imply that the after-ripening requirements of the seed they were handling might be more rigorous than the requirements of seeds grown in Italy. They pointed out that seeds of certain southern plants can be after-ripened at 15°C. as well as at lower temperatures. The present work may be criticized on this basis, since 15°C. was used as the minimum night temperature in the greenhouse. Even if it is assumed that the rapidly germinating seeds could have been after-ripened by the short periodic exposure to a temperature of 15°C., it is apparent that the same conditions did not result in germination of those seeds which were grown under the climatic conditions of 1949 and 1950. Thus, however the condition of the seeds is defined, it is still true that the pre-harvest climate has exerted a differential quantitative influence upon the dormancy.


Germination studies have been made of ten hybrid rose seed populations over a five-year period, and the results compared to variations in climate occurring during the 30-day and 60-day periods preceding harvest. Controlled hand pollinations and subsequent growth of fruit and seed were carried out in the field. Year-to-year environmental or circumstantial variations amenable to control in the field were reduced to a minimum, leaving climate as the major variable. The dormancy of hybrid rose seeds, and the consequent germination pattern in time, may vary widely between years, and low-temperature treatment is not invariably prerequisite to extensive germination. Good correlation is indicated 'between germination and preharvest climate, particularly with the mean of the average daily temperatures for the 30-day period preceding harvest. Aseptic culture of excised embryonic plants provided high uniform germination for all five years of the study, indicating that the embryo itself is generally complete and that its potential germinability was not influenced by the climatic variation encountered. The validity of the relationship between pre-harvest climate and seed dormancy is further substantiated by the results of experiments in which seed-bearing plants were transferred to a green-house cubicle in which they were provided high temperature and light. The subsequent germination of seeds from these plants indicates a marked reduction in dormancy as compared to field-grown controls.