The Korean Journal of Crop Science. September 2019. 213~224
https://doi.org/10.7740/kjcs.2019.64.3.213

# MAIN

• Introduction

• MATERIALS AND METHODS

•   Experimental materials and treatment conditions

•   Analytical and Statistical Methods

• RESULTS

•   Temperature-dependent variation in heading date by growth stage

•   Temperature-dependent variation in growth rate by growth stage

•   Major temperature parameters influencing the heading date by growth stage

• DISCUSSION

•   Effect of temperature on heading during the photo- sensitive period

•   Discrepancy between heading and elongation response to temperature

•   Daily temperature difference affect heading response during photo-sensitive period

• CONCLUSIONS

## MATERIALS AND METHODS

Experimental materials and treatment conditions

For our experiment, we used six rice (Oryza sativa) cultivars (temperate japonica type), two each for early maturing (Odae & Ungwang), medium maturing (Gopum & Daebo), and mid-late maturing (Sanuri & Shindongjin) ecotypes. For all cultivars, 20-day-old seedlings were transplanted on the same day (May 30) in 2017 and 2018 using 1/5000а Wagner pots, with three plants per pot. As a fertilizer, we used a composite slow-release fertilizer based on 9 kg/10a nitrogen, 4.5 kg/10a phosphate, and 5.7 kg/10a potassium at the area ratio corresponding to three plants (planting distance: 30 × 14 cm) instead of the pot area. Seedlings were sown and grown in 406-cell seedling trays. The average number of leaves on the transplantation day (May 30) was 3.9. Temperature treatment was applied to three groups of freshly transplanted rice plants in thermo-regulated glass rooms of the artificial weather facility of the National Institute of Crop Science: the control group was treated with temperatures (hourly automatic temperature control) varying at 9-day intervals in each treatment period based on the 15-year (2001–2016) average temperature in Jeonju area; the low- and high-temperature groups were treated with temperatures 3°C lower or higher than those applied to the control group (Fig. 1). Except for the temperature treatment period of nine days, the rice plants were grown in natural weather conditions.

##### Fig. 1.

Change in weather condition by days after transplanting. (A) Day length. (B) Average temperature. (C) Minimum temperature. (D) Maximum temperature.

Analytical and Statistical Methods

For the heading date analysis, “heading date” was defined as the number of days from transplanting to the appearance of the first spikelet, with nine replicates for each treatment group, as counted between 13:00 h and 14:00 h on a daily basis. The importance index of temperature on heading response in each growth stage was calculated through the following steps: (1) the temperature-dependent heading date variation at 9-day intervals was checked, (2) a value of +1 or –1 was assigned to one-day delay or advance, respectively, compared with the control group, (3) the heading date variation in response to a difference of 1°C in the daily average temperature during each treatment period was summed, (4) the total value was divided by the extent of variation in the corresponding period, (5) the value was expressed as a percentage. The vegetative growth phase was divided, based on the growth response of the control group grown in natural weather conditions, into the rooting stage (from transplanting to the emergence of the first leaf; ~9 days) and the early tillering stage (from the primary tillering stage, with the emergence of the primary tiller, to the secondary tillering stage with the emergence of the secondary tiller from the first node position; 7.5 leaves, ~9 days). The reproductive growth phase is generally considered to be completed in 35 days but varies depending on the criteria for determining the heading date (Vergara et al., 1965; Ikeda, 1975). In our experiment, given that the heading date was determined as the time when 50% of the panicles became exserted from the boot, we calculated panicle initiation, spikelet differentiation, and meiosis stages to be 30, 24, and 15 days prior to heading, respectively (Matsubayashi, 1963; Katayama, 1971).

The temperature-dependent variation in plant height prior to heading was computed using the following logistic model (Black & Leff, 1983):

 $$E=E_{min}+\frac{E_{min}+E_{max}}{1+(t/t_h)^{-Er}}$$

where E is plant height (cm), t is the number of days up to heading, Er is the curve pattern, which determines the rate of elongation reduction, Emax and Emin are the maximum and minimum elongation rate, respectively, and th is the time when 50% of Emax is reached. Er, Emax, Emin, andth are coefficients determined by nonlinear regression analysis, which was performed using SigmaPlot 11.1.

Results of each experiment were analyzed using the R Statistical Package (Version 3.2.2), with the differences considered to have statistical significance at p < 0.05. Random forest analysis was performed using the R-package randomForest and caret. The values of mtry and ntree, which determine the number of branches for explanatory variables and each node of the decision tree model, were set to those yielding the lowest OOB errors as follows, using tune RF and gridsearch: for the early maturing cultivars, mtry = 4, OOB = 0.93, and ntree = 30000; for medium maturing cultivars, mtry = 4, OOB = 2.15, and ntree = 10000; for the mid-late maturing cultivars, mtry = 5, OOB = 3.08, and ntree = 10000.

## RESULTS

Temperature-dependent variation in heading date by growth stage

As shown in Fig. 2, the influence of temperature on heading date varied by growth stage. In all ecological types, increase in temperature nonlinearly shortened the heading date in some growing stages, but there were also stages where no significant changes were observed, especially in the period between 36 and 27, 45 and 27, and 51 and 24 days before heading for early maturing, medium maturing, and mid-late maturing cultivars, respectively. Table 1 presents the magnitude of temperature effect on the heading date in different treatment periods as the difference in the days to heading induced by 1°C increase in the daily average temperature. The magnitude of temperature-dependent variation in heading date increased with an increase in the length of growth period, i.e., in the order of early maturing, medium maturing, and mid-late maturing ecotypes. The same trend was observed in the growth stages nonresponsive to temperature change. The growth period of rice plants is divided according to the days before heading: rooting, early tillering, panicle initiation, spikelet differentiation, and meiosis stages (Matsubayashi, 1963; Katayama, 1971). Fig. 3 illustrates the importance index in each growth stage. In all cultivars and ecotypes, the highest importance index was demonstrated by the rooting stage, followed by meiosis, early tillering, spikelet differentiation, and panicle initiation stages. In particular, during the photo-sensitive period (PSP), namely around the panicle formation time, the effect of temperature change on heading response was greatly reduced, which supports the finding of Ahn and Vergara (1969) that the effect of temperature was insignificant during PSP.

##### Fig. 2.

Change in heading date according to temperature variation by days before heading. (A) Early maturing type. (B) Medium maturing type. (C) Mid-late maturing type.

Table 1. Effect of degree of temperature on heading by days before heading.

 Days before heading (Days after transplanting) Early maturing (Days) Days before heading (Days after transplanting) Medium maturing (Days) Days before heading (Days after transplanting) Mid-late maturing (Days) 54 (1) 0.046 63 (1) 0.074 69 (1) 0.079 45 (9) 0.019 54 (9) 0.037 60 (9) 0.051 36 (18) 0.005 45 (18) 0.000 51 (18) 0.000 27 (27) 0.009 36 (27) 0.000 42 (27) 0.000 18 (36) 0.019 27 (36) 0.019 33 (36) 0.000 9 (45) 0.059 18 (45) 0.046 24 (45) 0.046 9 (54) 0.077 15 (54) 0.051 6 (63) 0.090 Total 0.156 Total 0.253 Total 0.317

- Effect of degree of temperature: Changes in heading date by 1°C of daily average temperature

##### Fig. 3.

Degree of heading response can be changed by 1 degree of daily average temperature at different growth stages. RT: Rooting stage; ET: Early tillering stage; PI: Panicle initiation; SD: Spikelet differentiation; MS: Meiosis stage; HD: Heading date.

Temperature-dependent variation in growth rate by growth stage

Fig. 4 shows the temperature-dependent variation in plant height growth rate by growth stage. No significant inter- ecotype differences were observed in the plant height growth rate after the 9-day temperature treatment in each growth stage. The average values (± SD) of plant height growth during the entire growth period in the low, normal, and high temperature conditions were 6.5 (± 1.2), 12.6 (± 2.1), and 17.3 (± 2.3) cm, respectively. This difference in growth response was maintained at similar levels until a specific period and sharply decreased after the peak growth period (Th in Table 2), which appeared earlier in the high- temperature treatment group. Contrary to the decrease in the effect of temperature on heading response around PSP, the effect of temperature on growth rate was maintained during PSP and began to decrease after the spikelet differentiation stage. Whereas, the correlation analysis between heading date and post-heading growth components, such as clum length, panicle length, number of spikelets per panicle, and number of panicles per plant, revealed a negative correlation only between the number of panicles and heading date (Fig. 5).

##### Fig. 4.

Degree of elongation response by days after transplanting with the logistic model. (A) Early maturing type. (B) Medium maturing type. (C) Mid-late maturing type.

Table 2. Parameters of the logistic function used to describe elongation response in relation to the days after transplanting under three temperature regimes during the growth period from transplanting to heading.

 Ecotype Treatment Emin Emax Th Er R2 Early Maturing Low-temperature 4.20 (0.05) 8.02 (0.02) 32.49 (0.25) 13.79 (0.86) 0.99** Control 5.53 (1.10) 17.44 (0.54) 31.58 (1.55) 12.11 (3.82) 0.97* High-temperature 3.36 (1.84) 24.89 (0.97) 30.32 (1.67) 14.35 (5.80) 0.97* Medium Maturing Low-temperature 2.86 (0.44) 7.54 (0.20) 39.95 (1.75) 17.62 (6.59) 0.96** Control 1.21 (0.78) 15.58 (0.32) 39.45 (0.89) 13.85 (2.73) 0.98** High-temperature 0.67 (2.84) 23.31 (1.01) 38.09 (1.89) 9.86 (4.44) 0.96** Mid-late Maturing Low-temperature 2.46 (0.48) 8.13 (0.22) 39.28 (1.59) 17.27 (6.66) 0.97** Control 1.58 (0.89) 17.51 (0.39) 39.20 (0.96) 15.22 (3.94) 0.98** High-temperature 0.25 (2.02) 25.85 (0.66) 38.94 (1.21) 10.97 (2.64) 0.98**

- Values in parentheses are standard errors
- "*" and "**" represent significance at the 0.05 and 0.01 probability levels, respectively.

##### Fig. 5.

Negative correlation between heading date and number of panicles per plant in all ecotypes.

Major temperature parameters influencing the heading date by growth stage

Table 3. Correlation coefficient between heading date and weather factors by temperature treatment at different growth stages.

 Weather Factors (DBH, DAT) Early maturing (Days) Weather Factors (DBH, DAT) Medium maturing (Days) Weather Factors (DBH, DAT) Mid-late maturing (Days) 1st-MXT (54, 1) -0.631** 1st-MXT (63, 1) -0.718** 1st-MXT (69, 1) -0.687** 6th-DTD (9, 45) -0.581** 1st-MNT (63, 1) -0.641** WP-AVT -0.633** 1st-AVT (54, 1) -0.561** 1st-AVT (63, 1) -0.639** 1st-DTD (69, 1) -0.623** 6th-MXT (9, 45) -0.560** 6th-MXT (18, 45) -0.634** 6th-MXT (24, 45) -0.626** 1st-MNT (54, 1) -0.557** 7th-MXT (9, 54) -0.633** 7th-MXT (15, 54) -0.615**

- "*" and "**" represent significance at the 0.05 and 0.01 probability levels, respectively.
- MNT: Minimum temperature, AVT: Average temperature, MXT: Maximum temperature, DTD: Daily temperature difference, WP: Whole growth period. DBH: Days before heading. DAT: Days after transplanting
- Information on the relationship between temperature treatment periods and growth stages can be found in Materials and Methods, Table 1.

##### Fig. 6.

Importance index (% increase MSE) of weather factors influencing heading through random forest analysis. A: Early maturing type. B: Medium maturing type. C: Mid-late maturing type. MNT: Minimum temperature, AVT: Average temperature, MXT: Maximum temperature, DTD: Daily temperature difference, WP: Whole growth period. Information on the relationship between temperature treatment periods and growth stages can be found in Materials and Methods, Table 1.

##### Fig. 7.

Partial dependence plot of the major factors affecting heading date in the early maturing type.

##### Fig. 8.

Partial dependence plot of the major factors affecting heading date in the medium maturing type.

## DISCUSSION

Effect of temperature on heading during the photo- sensitive period

Discrepancy between heading and elongation response to temperature

Unlike heading response, degree of elongation response increased with the increase in temperature until before the spikelet differentiation stage regardless of the growth stage, and no temperature-dependent difference in growth rate was observed afterwards (Fig. 4). In particular, in contrast to heading response, elongation response sensitivity to temperature was observed in PSP treatment. These results allow the assumption that vegetative growth rate can be predicted in temperature conditions within a specific range but predicting heading response based on them may be prone to error. From the finding that individuals with a higher number of panicles per plant showed an accelerated heading date (Fig. 5), it can be inferred that individuals that grow faster until the photoperiod-induced sensitivity to short daylength produce a larger number of tillers, resulting in faster heading. This can also be inferred from the report of Tsuji et al. (2015) that Hd3a, which regulates heading response, promotes tillering.

Daily temperature difference affect heading response during photo-sensitive period

A comparison of the average heading date between treatment periods revealed that while temperature treatment does not induce any significant differences in heading response, there are periods associated with a faster average heading date, namely the 4th treatment period in the medium maturing ecotype and the 4th and 5th treatment periods in the mid-late maturing ecotype (Fig. 2). In the results of random forest analysis as well, the importance index of daily temperature difference in the 4th treatment was the highest. This may be explained by the sudden decrease in daily temperature difference in the natural weather conditions due to the rise in minimum temperature in the 4th treatment period shortly after the summer solstice, whereas the daily temperature difference was maintained at 10°C in the artificial weather facility (Fig. 9). Comparing the growth stages among the maturing ecotypes, both the 4th and 5th treatment periods fell after panicle initiation in the early maturing cultivars, ~9 days before (in PSP) and after panicle initiation, respectively, in the medium maturing cultivars, and ~15 and ~6 days before panicle initiation (in PSP), respectively, in the mid-late maturing cultivars (Fig. 10). From these results, it can be inferred that daily average temperature has little impact on heading response during PSP prior to panicle initiation in normal growth conditions, but the nighttime minimum temperature and daily temperature difference affect heading response during PSP. Growth response to daily temperature difference over the entire growth period was demonstrated in the experiment conducted by Yin et al. (1997). Also, considering the experimental finding by Shibata et al. (1969) that the daytime temperature 29 and 32 days before heading had an insignificant impact on heading, but the nighttime temperature greatly influenced it, there is a need to carry out an in-depth study on the effect of daily temperature difference on heading response in PSP.

##### Fig. 9.

Change in daily temperature difference by days after transplanting under natural and experimental weather conditions. 4th, 5th: Days before heading [Early maturing: 27, 18; Medium maturing: 36, 27; Mid-late maturing: 42, 33].

##### Fig. 10.

Growth stage (days before panicle initiation) of each ecotype in the fourth and fifth treatment period.

## CONCLUSIONS

The following gives an overview of the key results regarding the temperature-dependent heading response of rice by growth stage and related explanations and interpretations:

1. The effect of temperature on heading response varies depending on the growth stage. It is the highest in the rooting stage, followed by meiosis, early tillering, spikelet differentiation, and panicle initiation stages; in particular, the effect of temperature in the photo-sensitive period (PSP) was verified to be negligible.

2. The analysis results for the temperature parameters influencing heading by growth stage revealed the change in daily temperature difference in PSP as one of the main factors, in addition to the temperature in the rooting and meiosis stages, and an in-depth study of this aspect is considered necessary.

3. Comparing the temperature-dependent growth response by growth stage, the temperature-dependent difference in growth response was maintained up to the panicle initiation period, but temperature treatment had little effect on growth response afterwards, which did not coincide with heading response.

## Acknowledgements

We thank all those who supported us in carrying out this research project. This paper presents the partial results of a study (title: Variation in heading date dependent upon temperature variations in different vegetative growth stages; project number: PJ01248801) funded by the grant of the research project sponsored by the South Korean Rural Development Administration.

## References

1

Ahn, S. B. and V. S. Vergara. 1969. Studies on responses of the rice plant to photoperiod III. response of korean varieties. Korean Journal of Crop Science 5 : 45-49.

2

Bernier, G. and C. Perilleux. 2005. A physiological overview of the genetics of flowering time control. Plant Biotechnol. J. 3 : 3-16.

10.1111/j.1467-7652.2004.00114.x17168895
3

Black, J. W. and P. Leff. 1983. Operational models of pharamacological agonism. Proc. R, Soc. Lond. B 220, 141-162.

10.1098/rspb.1983.00936141562
4

Breiman, L. 2001. Random forests. Machine learning 45 : 5-32.

10.1023/A:1010933404324
5

Collinson, S. T., R. H. Ellis, R. J. Summerfield, and E. H. Roberts. 1992 Durations of the photoperiod-sensitive and photoperiod-insensitive phases of development to flowering in four cultivars of rice (Oryza sativa L.). Annals of Botany 70 : 339-346.

10.1093/oxfordjournals.aob.a088483
6

Corbesier, L., C. Vincent, and S. Jang. 2007. FT protein movement contributes to long-distance signaling in floral induction of Arabidopsis. Science 316 : 1030-1033.

10.1126/science.114175217446353
7

Dua, A. B., F. W. T. Penning de Vires, and D. V. Seshu. 1990. Simulation to support evaluation of the production potential of rice varieties in tropical climates. Transactions of the American Society of Agricultural Engineers 33 : 1185-1194.

10.13031/2013.31456
8

Haniu, Y., H. Chujo, and S. Yhosida. 1983. Effect of air temperature on floral induction by short day in rice plants. Japanese Journal of Crop Science 52 : 135-142.

10.1626/jcs.52.135
9

Ikeda, K. 1975. Studies on initiation and subsequent development of floral bud in rice plants. III. Appearance of photoperiodic sensitivity and its changes with plant age. Bull. Fac. Agri. Mie Univ. 48 : 1-8.

10

Katayama, Y. 1971. Photoperiodism in the genus Oryza. VII. Mem. Fac. Agric. Kagoshima Univ. 8 : 299-320.

11

Kotak, S., J. Larkindale, U. Lee, P. V. Koskull-Dőrling, E. Vierling, and K. D. Scharf. 2007. Complexity of the heat stress response in plants. Current Opinion in Plant Biology 10 : 310-316.

10.1016/j.pbi.2007.04.01117482504
12

Komiya, R., A. Ikegami, S. Tamaki, S. Yokoi, and K. Shimamoto. 2008. Hd3a and RFT1 are essential for flowering in rice. Development 135 : 767-774.

10.1242/dev.00863118223202
13

Komiya, R., S. Yokoi, and K. Shimamoto. 2009. A gene network for long-day flowering activates RFT1 encoding a mobile flowering signal in rice. Development 136 : 3443- 3450.

10.1242/dev.04017019762423
14

Kovi, M. R., G. Sablok, X. Bai, M. Wendell, O. A. Rognli, H. H. Yu, and Y. Z Xing. 2013. Expression patterns of photoperiod and temperature regulated heading date genes in Oryza sativa. Computational Biology and Chemistry 45 : 36-41.

10.1016/j.compbiolchem.2013.03.00423688619
15

Lee, J. H., S. H. Park, and J. H. A, 2012. Functional conservation and diversification between rice osmads22/ osmads55 and arabiopsis svp protein. Plant Science 185 : 97-104.

10.1016/j.plantsci.2011.09.00322325870
16

Loomis, R. S. and D. J. Connor. 1992. Crop ecology: productivity and management in agricultural systems. Cambridge University Press Cambridge : 104-128.

10.1017/CBO9781139170161
17

Luan, W., H. Chen, Y. Fu, H. Si, W. Peng, S. Song, W. Liu, G. Hu, Z. Sun, D. Xie, and C. Sun. 2009. The effect of the cross talk between photoperiod and temperature on the heading-date in rice. PLoSONE 4, e5891. doi.:10.1371/ journal.pone.0005891.

10.1371/journal.pone.000589119521518PMC2690821
18

Matsubayashi, M. 1963. Theory and practice of growing rice. Min of Agr. and Fores. Tokyo. Japan 3 : 77-90.

19

Mimoto, H., M. Yanase, and H. Chujo. 1989. Varietal differences of juveninile phase duration and photoperiodic sensitivity in paddy rice. Japanese Journal of Crop Science 58 : 628-634.

10.1626/jcs.58.628
20

Nishida, H., Y. Okumoto, H. Nakagawa, K. Ichitani, H. Inoue, T. Tnisaka. 2001. Analysis of tester lines for rice (Oryza sativa L.) heading-time genes using reciprocal photoperiodic transfer treatments. Annals of Botany 88 : 527-536.

10.1006/anbo.2001.1490
21

Roberts, E. H. and R. J. Summerfield. 1987. Measurement and prediction of flowering in annual crops. In: Atherton JG, ed. Manipulation of flowering. London: Butterworths, 17-50.

10.1016/B978-0-407-00570-9.50007-7PMC1338742
22

Satake, T. and S. Yoshida. 1978. High temperature-induced sterility in indica rices at flowering. Japanese Journal of Crop Science 47 : 6-17.

10.1626/jcs.47.6
23

Shi, P. H., Y. Zhu, L. Tang, J. L. Chen, T. Sun, W. X. Cao, and Y. C. Tian. 2016. Differential effects of temperature and duration of heat stress during anthesis and grain filling stages in rice. Environmental and Experimental Botany 132 : 28-41.

10.1016/j.envexpbot.2016.08.006
24

Song, Y. and W. Luan. 2012. Molecular regulatory network of flowering by photoperiod and temperature in rice. Rice Science 19 : 169-176.

10.1016/S1672-6308(12)60037-7
25

Song, Y., Z. Gao, and W. Luan. 2012. Interaction between temperature and photoperiod in regulation of flowering time in rice. Science China Life Sciences 55 : 214-249.

10.1007/s11427-012-4300-422527521
26

Song, Y. H., S. Ito, and T. Imaizumi. 2013. Flowering time regulation: photoperiod and temperature-sensing in leaves. Trends in Plant Science 18 : 575-583.

10.1016/j.tplants.2013.05.00323790253PMC3796012
27

Tamaki, S., S. Matsuo, H. L. Wong, S. Yokoi, and K. Shimamoto. 2007. Hd3a protein is a mobile flowering signal in rice. Science 316 : 1033-1036.

10.1126/science.114175317446351
28

Vergara, B. S., S. Puranabhavung, and R. Lilis. 1965. Factors determing the growth duration of rice varieties. Phyton 22 : 177-185.

29

Vergara, B. S. and T. T. Chang/ 1985. The flowering response of the rice plant to photoperiod. 4th edn, Los Banos, Philippines: IRRI.

30

Wang, J. Y. 1960. A critique of the heat unit approach to plant response studies. Ecology 41 : 785-790.

10.2307/1931815
31

Wang, Y., L. Wang, J. Zhou, S. Hu, H. Chen, J. Xiang, Y. Zhang, Y. Zeng, Q. Shi, D. Zhu, and Y. Zhang.2019. Research progress on heat stress of rice at flowering stage. Rice Science 26 : 1-10.

10.1016/j.rsci.2018.06.009
32

Yamamoto, T., H. Lin, T. Sasaki, and M. Yano. 2000. Identification of heading date quantitative trait locus Hd6 and characterization of its epistatic interactions with Hd2 in rice using advanced backcross progeny. Genetics 154 : 885-891.

33

Yang, Z., L. Jin, H. Zhu, S. Wang, G. Zhang, and G. Liu. 2018. Analysis of epistasis among QTLs on heading date based on single segment substitution lines in rice. Scientific Reports 8, 3059.

10.1038/s41598-018-20690-w29449579PMC5814450
34

Yano, M., Y. Harushima, Y. Nagamura, N. Kurata, Y. Minobe, and T. Sasaki. 1997. Identification of quantitative trait loci controlling heading date in rice using a high-density linkage map. Theoretical and Applied Genetics 95 : 1025-1032.

10.1007/s001220050658
35

Yin, X., M. J. Kropff, and J. Goudriaan. 1996. Differential effects of day and night temperature on development to flowering in rice. Annals of Botany 77 : 203-213.

10.1006/anbo.1996.0024
36

Yin, X., M. J. Kropff, and J. Goudriaan. 1997. Changes in temperature sensitivity of development from sowing to flowering in rice. Crop Science 37 : 1787-1794.

10.2135/cropsci1997.0011183X003700060021x
37

Yin, X., M. J. Kropff, T. Horie, H. Nakagawa, H. G. S. Centeno, D. Zhu, and J. Goudriaan. 1997. A model for photothermal responses of flowering in rice I. Model description and parameterization. Field Crops Research 51, 189-200.

10.1016/S0378-4290(96)03456-9
38

Yoshida, S. 1981. Fundamentals of rice crop science. IRRI, Los Banos, Laguna, Philippines, 269.

39

Yoshida, S. and Y. Hanyu. 1964. Critical daylength for rice plants in realtion to temperature [in japanese]. Proc. Kinki Symp. Plant Breed. Crop Sci. Soc. 9 : 34-36.