Effects of temperature sum on vitamin C concentration and yield of sea buckthorn (Hippophae rhamnoides ) fruit: optimal time of fruit harvest Yingmou Yao Yao, Y. 1993. Effects of temperature sum on vitamin C concentration and yield of sea buckthorn ( Hippophaerhamnoides) fruit: optimal time of fruit harvest. Agric. Sci. Finl. 2; 497-505. (Dept. Plant Biology, FIN-00014 University of Helsinki,Finland.) To investigate the effects of temperature sum on vitamin C concentration (Vc ), yield and maturity of sea buckthorn fruit (Hippophae rhamnoides L.) and to predict the optimal harvest time, berries were collected from eight genotypes at an interval of about one week from August 16 to December 2. Maturity was visually observed, berry weight measured and Vc determined. Berries matured at 1165-1316 degree-days (d.d.). Vc reached maximum at about 1229 d.d., while fruit size and yield reached maximum at 1380 d.d.. Mathematical models of polynomial equations were highly significant for predicting the effects of temperature sum on Vc , maturity and fruit yield. Optimal harvest time for maximizing Vc, yield or economic income could be determined according to differential equations. Great variations in Vc , fruit maturity and fruit size suggested good opportunities for selection and breeding. Low rank correlations in vitamin C concentration during fruit maturity, however, call for special attention in selection and breeding. Key words; fruit size, maturity, mathematical models, breeding Introduction Sea buckthorn, Hippophae rhamnoides, is widely distributed on the Eurasian continent. Its agricul- tural, nutrient, medical and ornamental value makes the plant a very promising subject for do- mestication. The nutrientand medical values have commanded great attention during the last decade. The birth of the special Chinese journal "Hippo- phae" in 1988 has provided researchers with infor- mation on the importance of the plant in China. The berries of sea buckthorn are among the rich- est sources of vitamin C in edible fruits. There are a number ofreports on the vitaminC concentrations (V c ) of H. rhamnoides berries. Results have revealed a large variation ofVc among subspecies, populations and genotypes of H. rhamnoides (Tian 1985, Li et al. 1988, Plekhanova 1988, Wang et al. 1990, Zhao et al. 1991, Wahlberg and Jeppsson 1990, 1992). Yao et al. (1992) have in- vestigated the variation of Vc between and within natural sea buckthorn populations in Finland. All these reports indicate good opportunities for selec- tion and breeding. Nonetheless, Rousi and Aulin (1977), Yang et al. (1988), Liu et al. (1990), Wahlberg and Jeppsson (1990) found a signific- ant decline of Vc during fruit maturity. Their find- ings imply that harvest time is very critical for securing a high Vc in sea buckthorn berries. These studies showed the relations between Vc and the 497 Agric. Sei. Fin!. 2 (1993) https://www.c-info.fi/en/info/?token=J0s5yBbgjvWx-hJ4.oFmRsY2gr4UmCGkSFG-35g.578eFahOcx99JThrnkGfEaNZea3kVnmcxCX0gQ8Bx--ZRrjy2Cf3Z5x1lWb35fNIeJ8XXsrHdYCpPC16BwNBQVWuiIXtA8qGqk7w3B-fHKgT0njNSNwQV9kH0gMgJOVlhxDTJwiyxjF86aw2rT6UjlGnkrJamU2w08GeMIFx4hBtEbnjWyJXys0sZcwt2fxWDE2kLGTsJxFjQQjhnVsivXmdIh4kVrQ6Lirb7uoCiL7BpjIwkIwkojUU9B2n78cfq9M calendar dates of fruit collection. However, many studies have shown that dates of biological devel- opmental stages vary greatly from year to year due to yearly climatic fluctuations. Temperature sum or heat sum Ts degree-days (d.d.), particularly the effective temperature sum (>5°C), has been widely used for observing the phenology of horticultural, agricultural and silvicultural plants. Results have constantly indic- ated that T s is a better predicator than calendar date for budbreak, flowering, fruit setting, harvest time and yield (Daniel and Bajtay 1984,Kristensen et al. 1987, Rysava and Poruba 1987, Hari and Häkkinen 1991). The purpose of the present study was to invest- igate the effects of temperature sum Ts on fruit maturity, vitaminC concentrationVc and fruit yield in sea buckthorn. The study also addressedrelation- ships among these variables. A model is proposed to predict the optimal time of harvesting sea buck- thorn berries. In addition, ranking and selection of genotypes for breeding are discussed. Material and methods From August 16 to December 2, 1990, berry samples were collected separately from eight indi- vidual bushes of a plantation growing in Helsinki harbour. Collections were made at approximately weekly intervals, except for the last three collec- tions, which were taken at intervals oftwo and three weeks. The plantation was of Danish origin and 18 years old, consisting of about 400 bushes. Bushes were selected at a regular interval in the plantation, and the collected sample was represented by eight genotypes. When sea buckthorn berries mature, their colour usually turns yellow, orange or red depending on the genotype. These colours are normally taken to indicate fruit maturity. In the present study, the process of fruit maturity of sea buckthorn berries was observed on the basis of colour changes. The percentage of matured berries (Mp ) on each bush, i.e. the proportion of yellow, orange or red berries on each bush, was recorded according to visual observation during collection at the site. This per- centage Mp was used as a quantitative measure of the degree of fruit maturity on a bush or group of bushes. When Mp was 100%, the bush had reached full maturity. After each collection, berry weight (weight/100 berries) was measured immediately in three repli- cations, except for bush 8 which bore too few ber- ries to allow this measurement. The samples were then kept at -20°C until determination of Vc . Vc was determined by high-performance liquid chromatography (HPLC) on the collection day or the following day according to the method devel- oped by Vuorela et al. (1986). To avoid any loss of vitamin C by breaking of berries, approximately 10 g of unbroken berries were exactly weighed from each bush. After weighing, each sample was immediately immersed in 20 ml of extractant solu- tion, then homogenized and centrifuged. In order to make the sample concentration fall within the cali- bration range, 1 ml of sample solution was diluted with 2 ml of extractant. A 10 (al aliquot of this diluted sample solution was used in the HPLC sys- tem for analysis. An external standard was used after every four samples had been chroma- tographed. 9The calibration curve (r = 0.998) was based on six replications for each amount of injection. The computation ofVc was based on the exact weightof each berry sample (including seed weight). Follow- ing the convention, Vc was expressed as mg% (mil- ligrams of vitamin C per 100 grams of berries). The above method detects only L-ascorbic acid. In addition, what is usually called vitamin C also includes the oxidized form ofL-ascorbic acid (de- hydroascorbic acid). However, its amount seems to be generally small in mature fruits (Mapson 1970) and is ignored in most analyses (Yao et al. 1992). The effective temperature sum (over 5°C), sim- ply called temperature sum T s, was calculated based on the daily mean temperature observed at the Kaisaniemi Weather Station (Ilmatieteen laitos 1990), one kilometer from the plantation. Ts reached 1419.4 d.d. on November 2 and then ceased to increase any further. SAS programs RSQUARE and GLM (SAS Institute 1985) were used for calculation and data analysis including modelling. 498 Agric. Sei. Fin!. 2 (1993) Results Effects of temperature sum on fruit maturity Observations showed that the sea buckthorn berries changed from green to yellowish green, then to greenish yellow and finally to yellow, orangeor red during their maturity process. There was a large variation in fruit maturity among genotypes (Fig. 1).Bush 8, the earliest one, was 64 d.d. (one week) earlier than bush 3, 128 d.d. (two weeks) earlier than bushes 4 and 6, and 151 d.d. (three weeks) earlier than the rest. Recent years’ phenological observations have confirmed that bush 8 is one ofa few genotypes characterized by early flowering, budbreak, growth cessation, maturity and leaf se- cession. The results indicated that there are good opportunities for selection and breeding for early maturity. Effect of temperature sum on vitamin C concentration The effects of temperature sum T s on vitamin C concentration Vc are shown in Fig. 2. There were two peaks in Vc for all bushes, except that there was only one poorly defined peak for bush 8. The first peak was distinct and high for all bushes except that it was flat for bushes 7 and 8. The second peak was also obvious though it was much less pronounced than the first one. The first peak of Vc appeared at 1229 d.d. (August 31) for all but for bush lat 1165 d.d. (August 24). Bushes 7 and 8 always ranked lowest, while bush 5 almost invariably ranked highest for Vc . The remaining five bushes varied considerably with collection time and frequently changed their ranks. This indicates that except for the genotypes with extreme values of Vc , selection of genotypes for Fig. 1.Effects of temperature sum on fruit maturity. Fig. 2. Effects of temperature sum on vitamin C concentra- tion. (Points for Nov. 11 and Dec. 2on the X axis given in time scale, see text). 499 Agric. Sei. Fin!. 2 (1993) high Vc based on one measurement at one time should be used with caution. In fact it would be difficult to separate genetic variance from experi- mental error variation in such genotypes. Effects of temperature sum on fruit yield Several years’ observation has shown that prema- ture drop and developmental failure of sea buck- thorn fruit occur only in the early stages of fruit development. A normally developing berry in July almost always proceeds towards maturity. There- fore the number ofberries on a bush after the end of July remains basically unchanged (a constant) and thereafter the fruit yield (Yf) of a given bush or plantation is determinedonly by the average size or weight of fruit. In this case the fruit weight or size is an index of fruit yield, and the effects of Ts on fruit weight or fruit size are equivalent to the effects of T s on fruit yield of a given bush or plantation. Therefore in the present study, the effects of Ts on fruit weight, fruit size and fruit yield are used inter- changeably whenever convenient. However, when different bushes or genotypes are compared, fruit weight or size is no longer a proportionate index of fruit yield. The effects of temperature sum T s on fruit yield Yf are presented in Fig. 3. The variationofYf of sea buckthorn with maturity appears as a curve with a single peak. With the increase of T s, Yf increased steadily until 1367 d.d. (September 30) when bushes 4 and 2 reached their maximum values. At 1380 d.d., bushes 3,5, 6, 7 and virtually the entire plantation reached their maximum yield, while bush 1 reached its maximum yield at 1398. The ranking of bushes based on the fruit weight was rather consistent in contrast to the rankings based on Vc . This suggests that genotypes with large berries will generally retain this characteristic regardless of collection time and that selection for genotypes with large berry size based on a single measurement in time is reliable. Relationships among vitamin C concentration, fruit yield, vitamin C yield and fruit maturity From the results (Figs. 1,2 and 3) we know that the maximum Vc appeared at 1229 d.d., 87 d.d. earlier than full maturity (1316 d.d) and 151 d.d. earlier than maximum fruit yield Yf (1380 d.d.). To enable a meaningful examination of the general relations among Vc, Yf and Mp, averaged values of Vc , Yf and Mp over the eight bushes were expressed as a percentage of their corresponding maximum values (Fig. 4). From 1075 d.d., the first collection, to 1229 d.d, V c , Yf and Mp all increased with Ts . At 1229 d.d., Vc reached its maximum, while Yf and Mp reached 77% of their maxima. Thereafter Vc started to de- crease, while the fruit yield Yf and fruit maturity Mp continued to increase untilMp reached 100% at 1316 d.d.. Beyond this Vc continued to decrease Fig. 3. Effects of temperature on fruit yield (fruit size). (Points for Nov. 11 and Dec. 2on the X axis given in time scale, see text). 500 Agric. Sei. Fin!. 2(1993) while Yf still kept increasing until Yf reached its maximum at 1380 d.d.. At this temperature sum Vc reached its minimum value, 54.3% of the maxi- mum. Again, Vc began to increase and Yf to de- crease untilVc approached its second peak, where- after both decreased sharply. It is interesting to note that between 1229 and 1418 d.d., Yf and V c were always negatively corre- lated (Fig. 4). One may wonder whether the amount of vitamin C in the berries was constant (maybe dynamically) during this period, and the variation in Vc caused only by growth in fruit size. To answer this question, a new measure, vitamin C yield Vy was calculated. V y was defined as V c multiplied by Yf, Vy = V cYf, i.e. total amount ofvitamin C in 100 berries (or a berry, ora bush). If the variation in Vc was caused only by an increase or decrease in fruit size, Vy should be constant during fruit maturity. However, this was not the case. Fig. 4 tells us that the amount of vitamin C in a berry Vy was not constant during the fruit maturity and post-maturity processes. Therefore the variationofVc was caused not only by fruit size Yf but also, and more import- antly, by the variationof total amount of vitaminC in a berry, i.e. the amount of vitamin C synthesized and decomposed in a berry during the fruit maturity and post-maturity processes. Modelling the effects of temperature sum on vitamin C concentration, fruit yield, vitamin C yield and maturity The results above showed that Mp , Vc , Yf and Vy were all functions of temperature sum Ts. Therefore mathematical models could be built to simulate and predict the effects ofT s on Mp , Vc , Yf and Vy, and furthermore to determine the optimal harvest time. Since the temperature sum showed no further in- crease after November 2 and the fruit of sea buck- thorn is rarely harvested later than the end of Octo- ber, only the data before the end of October were used for modelling and the resulting models were applicable to the same period. Mathematical models could be built for each bush or genotype if necessary. However, the fol- lowing mathematicalmodels in Table 1for simulat- ing and predicting the effects of Ts on Mp, Vc, Yf Fig. 4. Relationships among vitamin C concentration (V c ), fruit yield (Yf), fruit maturity (Mp ) and vitaminC yield (Vy). (Points for Nov. 11 and Dec. 2on the X axis given in time scale, see text). 501 Agric. Sei. Finl. 2 (1993) Table 1. Models for predicting fruit maturity, vitamin C concentration, fruit yield and vitamin C yield (P < 0.015 for all parameters in the equations). 1 Model R 2 F P 1 Mp = -313 + 314T S 2 Vc = 372051-1225143T S + 1505770T2 -818346T 3 + 165946T 3 3 Yf = -27081 + 88365T S -107706T 2 + 58159T 3-! 1734T* 4Vy = 92721-304691TS + 373578T 2-202492T 3 + 40950T3 0.978130.6 0.0014 0.98797.0 0.0001 0.989116.5 0.0001 0.97345.3 0.0004 5 Vc = 180607-594730T, + 730956T2-397255T 2 + 80557T* 6 Yf = -72799 + 237540T S -289533T 2 + 156340TS 3-31542T‘ 7 Vy = 157816-518579T, + 635794T2 -344605T 2 + 69687T? 0.98797.0 0.0001 0.989116.5 0.0001 0.97345.3 0.0004 1 Mp = percentage of matured berries, Ts = effective temperature sum, Vc = vitamin C concentration, Y, = fruit yield, V y = vitamin C yield, (see text for further explanation). and Vy are based on the average values of the eight bushes. The equations in Table 1 are divided into two groups. The upper group predicts the values of the variables Vc , Yf, Vy and Mp in the units defined in the text, and the lower group predicts the percent- age relative to their maximum values. In both cases Ts in the equations is the temperature sum divided by 1000. If one’s purpose is to know actual values, equations in the upper group are used. If the pur- pose is to know the value of a variable relative to its maximum or to determine the optimal harvest time, the lower groupof equations are more suitable. The R , F and P values indicated that all models listed in Table 1 have a good predictive power. If the power of polynomial equations was reduced to 2, R was still as high as 0.942 for Yf but dropped to 0.656 and 0.730 for Vc and V y, respectively. Predicting and determining optimal time of harvest With the above equations, the optimal time of har- vest, which maximized an objective, can be mathe- matically determined. This was done by taking dif- ferentials of above equations and solving for Ts from the differential equations, i.e. by solving the equation dF(Ts)/dTs = 0. The solutions for T s must be multiplied by 1000 to give the temperature sum. Optimal harvest timefor maximizing vitamin C concentration If the objective is to maximize Vc , vitamin C con- centration, from dVc/dTs = owe getT s = 1233 d.d., where Ts means the value of Ts which maximizes Vc . This was 83 d.d. (or about 2 weeks) earlier than the time of full maturity (1316 d.d.), or at the time corresponding to 77% of full maturity. From Figs. 2 and 4 and equations 2 and 5 we know that Vc increased and dropped sharply on either side of its maximum. This sensitivity of Vc to Ts suggests a narrow period ofoptimal time ofharvest for obtain- ing higher Vc . In other words, to obtain higher Vc, harvest must be carried out within a very short period of time. Optimal harvest timefor maximizing fruit yield If the objective is to obtain maximum fruit yield Yf, dYf/dTs = 0 gives the optimal harvest time T s = 1389 d.d. which was 156(39 days) and 73 d.d. (26 days) later than those for Vc (1233) and full matur- ity (1316), respectively. Figs. 3 and 4, and equa- tions 3 and 6 all show that Yf was insensitive to T s around its maximum value. This indicates a rather long optimal time of harvest for fruit yield Yf. Actually, harvest could be conducted for a period of 37 (62) d.d. and still ensured Yf over 99% (95%) of its maximum. Furthermore because of the slow in- crease of Ts in later September onwards, 37 (62) 502 Agric. Sd. Fin!. 2 (1993) d.d. ofT s was equal to about 2 (5) weeks ofharvest time. Optimal harvest timefor maximizing vitamin C yield If the objective is to obtain maximum vitaminC yield Vy, i.e. the purpose of harvest is to extract and to produce natural vitamin C, then dVy/dTs = 0 gives the solution of Ts = 1250 d.d. which was 17 d.d. (2 days) later and 66 d.d. (11 days) earlier than the time for Vc and fruit maturity. The harvest time could also be slightly longer for Vy because V y was not as sensitive as Vc to Ts after its maximum. Optimal harvest time for maximizing income I In the above discussion of optimization, the first consideration was to maximize nutrient concen- tration or quality and the subsequent considera- tions were to maximize yield. From the standpoint of commercial production the objective is to maximize the income or profit, i.e. to maximize the expression I = PYf whereI = income, P = price and Yf = fruit yield as before. If the price is a function of V c , i.e. P = P(V C), the equation becomes I = PYf = P(Vc )Yf. We already know that Vc and Yf are functions ofTs, so income lis also a function of Ts. To maximize I is to find out the solution for dl/dTs = 0. This could be done mathematically provided that the equation P = P(V C) was defined, i.e if the relation between the price and Vc were known. If the price Pis a linear function ofVc , i.e. P = a + pvc, then I = PYf =(a + pvc )Yf. If the price is constant regardless of Vc, thenP = O,P = a>o and I = aYf, and the T s maximizing fruit yield Yf will also maximize the income I. If the price is propor- tional to V c, then a =O,P> 0, P = pvc and I = pVcYf = pVy, and the T s maximizing vitamin C yield Vy will also maximize the income I. If neither a nor P is equal to zero, the Ts maximizing the income must lie between the two Ts values which maximize Vc and Yf. If ais relatively smaller, the solution ofTs is closer to the T s maximizing Vc , but if a is relatively larger it is closer to the T s max- imizing Yf. Discussion Many studies have revealed large variations of Vc in sea buckthorn berries among subspecies, popula- tions and genotypes, suggesting good opportunities for breeding and selection for high Vc . Variation of Vc during the fruit maturity process has, however, posed another important question: what is the opti- mal harvest time? Results of the present study showed that Vc of sea buckthorn berries reached a maximum at 1229 d.d., 87 d.d. earlier than full maturity (1316 d.d.) and 151 d.d. earlier than the maximum yield (1380 d.d.). At that time the fruit had attained 77% of full maturity and its yield was 77% of the maximum. When the berries reached full maturity at 1316 d.d., the fruit yield had reached 88% of its potential, while Vc had dropped to 74% of its maximum. When yield reached its maximum value at 1380 d.d., Vc dropped to its minimum, about half of its maximum value. Yang et al. (1988) found a peak and a valley of Vc during fruit maturity, similar to that observed in the present study. Liu et al. (1990) showed only a single peak of Vc, corresponding to the first peak of this study, because of their short study period. Rousi and Aulin (1977), Wahlberg and Jeppsson (1990) showed only a decline of Vc , cor- responding to the right part of the first peak in this study, because of their later beginning and short study period. Actually the peak and later rising of Vc could be found by carefully examining the scat- tering of WAHLBERG-JEPPSSON’s Fig. 3. Therefore the pattern of variation shown in this study appears common in sea buckthorn. The variation of fruit size during fruit maturity was similar to that found in other studies though they used date rather than temperature sum. Yang et al. (1988) indicated that the secondary rise ofVc was probably due to a decrease of fruit weight. The vitamin C yield Vy calculated in the present study showed that the increase and decrease of Vc during fruit maturity followed the total amount of vitamin C synthesized and decomposed. In rose hips, Vc also shows peak variation during fruit development (Uggla 1988). However, a decline ofVc is gener- ally found in black currant and other fruits during 503 Agric. Sei. Finl. 2 (1993) ripeness (HÅRDH 1964, SISTRUNK et al. 1983), simply because the peak appears at the immature stage before the analysis starts. The secondrising of Vc is probably a defending reaction of the postma- tured fruit, and the following sharp decrease might indicate the deterioration of fruit, similar to the respiration peak during the storage in apples, pears and other fruits. In the present study, the author developed math- ematical models based on temperature sum Ts pro- viding good predictions of fruit yield, vitamin C concentration and vitamin C yield. The optimal harvest time could be determined by the models. The best harvest time, however, depends on one’s objectives. Vitamin C concentration is only one of many nutrient and other quality factors. For com- mercial production the price offruit is a function of quality index (integrating all nutrient and other quality factors). If Vc is the only important factor, optimal harvest time can be determined for maxim- izing Vc, Vy, Yf or the economic income. Further- more, the models provide a flexible way of deter- mining optimal harvest time under one or more constraints of quality, weather or labour restric- tions. The present study using temperature sum as pre- dictor overcame some of the problems caused by using calendar date, and the models fitted the actual data very well. However, as the temperature sum reaches a ceiling value in the late autumn, further observations on a "time scale" must be given in calendar dates, as indicated in Figs. 2-4. At this stage offruit development changes may occur that are independent of the Ts-dependent "biological clock". The study showed large differences in fruit maturity among genotypes. The earliest genotype was 151 d.d. earlier than the latest ones. The results have also been confirmedby the observations made in recent years by the author. This suggests good opportunities for selection and breeding for early maturity, particularly useful under Finnish condi- tions. The relatively constant rank of genotypes based on fruit size should make it easy to breed for large berries, an important yield component, par- ticularly in developing an efficient harvesting tech- nique. The low rank correlation among genotypes based on Vc during fruit maturity calls for special attention in selection and breeding. In addition, the large variation of Vc during different collection times makes comparison of differentanalyses diffi- cult. Acknowledgements. The author wishes to thank Prof. P.M.A. Tigerstedt and Mr. P. Joy for valuable suggestions and for checking the English. Thanks are also extended to the Acad- emy of Finland for providing the grant. References Daniel, L. & Baitay, I. 1984. Flowering time prediction in sweetcorn varieties of the Vegetable Crops Research In- stitute. Zoldsegermesztesi Kutato Intezet Bulletinje 17: 5-12. (in Hungarian, English abstract). Härdh, J. E. 1964. Mustaherukan C-vitamiinipitoisuuteen vaikuttavista tekijöistä. Maatal. tiet. Aikak. 36: 14-21. Hari, P. & Häkkinen, R. 1991. The utilization of old pho- nological time series of budburst to compare models describing annual cycles of plants. Tree Physiology 8,3: 281-287. Ilmatieteen laitos 1990. Kuukausikatsaus Suomen ilmastoon (Monthly weather reports by the Finnish Meteorological Institute. Reports January-December. (in Finnish). Kristensen, S., Friis, E., Henriksen, K. & Mikkelsen, S. A. 1987. Application of temperature sums in the timing of production of crisp lettuce. Acta Horticulturae 198: 217- 225. Li,C., Ma,Z., Liu, D. & Zhang,W. 1988.A study of the loss rate of total acid and vitamin C of common sea buckthorn raw juice in different concentrated conditions. Hippo- phae 2: 28-29. (in Chinese). Liu, D., Zhang, W., Deng, X., Li, C, Zhang, H., Dou, Y. & Li, Y. 1990. Research on the dynamic changes in vit- amins of sea buckthorn fruit during its growing stage. Hippophae 3: 32-35. (in Chinese). Mapson, L.W. 1970. Vitamins in fruits. In: Hulme, A. C. (ed.). The Biochemistry of Fruits and Their Products I. London, p. 369-384. Plekhanova, M. N. 1988.Sea Buckthorn. 77 p. (in Russian). Rousi, A. & Aulin, H. 1977. Ascorbic acid content in rela- tion to ripeness in fruits of six Hippophae rhamnoides clones from Pyhäranta, SW Finland. Ann. Agric. Fenn. 16: 80-87. Rysava, B. & Poruba, J. 1987. Effect of number of days, heat sum, sunlight and precipitation during the sowing- flowering period in some approved grain maize hybrids. 504 Agric. Sei. Finl. 2 (1993) Pol’nohospodarstvo 33, 8: 724-731. (in Slovakian, Eng- lish abstract). SAS Institute Inc. 1985. SAS User’s Guide: Statistics. sth Ed. Cary, NC, SAS Institute Inc. 956 p. SisTRUNK, W. A. & Moore, J. N. 1983. Quality. In: Moore, J.N. & Janick J. (eds.). Methods in fruitbreeding. Purdue Uni., Indiana, p. 274293. Tian, H. 1985. Hippophae: cultivation and Development. Shaanxi Agric. Info. Press. 32 p. (in Chinese). Vuorela, H., Hinkkanen, R. & Hiltunen, R. 1986. Rapid determination of ascorbic acid in pharmaceuticals and biological samples by HPLC. Acta Pharm. Fenn. 95: 141-149. Ugcla, M. 1988. Framtagning av odlingssystem och sortma- terial för etablering av nyponodlingar. Abstract: Devel- opment of rose hip cultivars and growing techniques for establishment of plantations. Sveriges Lantbruksuniver- sitet Balsgård-Avdelningen för Hortikulturell Växt- förädling Verksamhetsberättelse 1986-1987p. 47-54. Wahlberg, K. & Jeppsson, N. 1990.Framtagning av sortma- terial och rationella odlingssystem för Havtom, Amnia, Lonicera och Sorbus. Abstract: Development ofcultivars and growing techniques for sea buckthorn, black choke- berry, Lonicera and Sorbus. Sveriges Lantbruksuniver- sitet Balsgård-Avdelningen för Hortikulturell Växt- förädling Verksamhetsberättelse 1988-1989. p. 80-93. & Jeppsson, N. 1992. Framtagning av sortmaterial och rationella odlingssystem för havtom, aronia, blåtry och rönn. Abstract: Development of cultivars and growing techniques for sea buckthorn, black chokeberry, honey- suckle and rowan. Sveriges Lantbruksuniversitet Bals- gård-Avdelningen för Hortikulturell Växtförädling Verk- samhetsberättelse 1990-1991.p. 86-100. Wang, G., Zhao, X., Chen, J. & Guo, J. 1990. A survey of fruit bearing laws of Chinese sea buckthorn plants. Hip- pophae 3: 19-22. (in Chinese). Yang, H., Wang, S. & Su, Y. 1988. A study on the dynamic variation of ascorbic acid in Hippophae. Hippophae 4: 41-44. (in Chinese). Yao, Y., Tigerstedt, P. M. A. & Joy, P. 1992. Variation of vitamin C concentration and character correlation be- tween and within natural sea buckthorn ( Hippophae rhamnoides L.) populations. Acta Agric. Scand. 42: 12- 17. Zhao, H. Zhu, C„ Gao, C„ Li, H., Liu, Z. & Sun, W. 1991. Geographical variations of fruit traits of the Chinese sea buckthorn and selection of provenances for fruit use. Hippophae 4: 15-18. (in Chinese). Manuscript received August 1993 Yingmou Yao Department of Plant Biology P.O. Box 27 FIN-00014 University of Helsinki, Finland SELOSTUS Lämpösumman vaikutus tyrnin marjasatoon ja marjojen C-vitamiinipitoisuuteen Yingmou Yao Helsingin yliopisto Lämpösumman vaikutusta tyrnin (Hippophaea rhamnoides) C-vitamiinipitoisuuteen, marjasatoon ja kypsyysasteeseen tutkittiin kahdeksalla eri genotyypillä. Optimikoijuuajankoh- dan määrittämiseksi koejäsenistä otettiin näytteitä viikon vä- lein elokuun puolivälistä joulukuun alkuun. Kerätyistä näyt- teistä määritettiin visuaalisesti kypsyysaste, punnittiin satoja määritettiin C-vitamiinipitoisuus. Marjat kypsyivät lämpö- summan ollessa 1165-1316 astetta. Maksimi C-vitamiinipi- toisuus havaittiin noin 1229asteen lämpösummalla, kun taas marjojen kokoja samalla maijasato olivat suurimmillaan läm- pösumman ollessa 1380astetta. Havaintojen pohjalta laadittiin matemaattinen malli, jonka avulla voidaan ennustaa lämpösumman vaikutus tarkasteltui- hin ominaisuuksiin. Vastaavasti sadonkorjuun optimaalinen aika C-vitamiinipitoisuuden, sadon ja taloudellisen tuloksen kannalta voitiin määrittää mallin pohjalta laaditun yhtälön avulla. Genotyyppien välinen suuri muuntelu tyrnin C-vita- miinipitoisuudessa, maljojen kypsyysasteessa ja marjojen koossa osoittavat tyrnin jalostukselle olevan hyvät mahdolli- suudet. 505 Agric. Sei. Fint. 2 (1993)