Rootstock effects on responses of potted ‘Smoothee Golden Delicious’ apple to soil-applied triazole growth inhibitors. II: Mineral nutrition and carbohydrate status

Scientia Horticulturae, 46 ( 1991 ) 75-88 75 Elsevier Science Publishers B.V., Amsterdam

Rootstock effects on responses of potted 'Smoothee Golden Delicious' apple to soil-

applied triazole growth inhibitors. II: Mineral nutrition and carbohydrate status*

Jeffrey K. Zeller l, Fenton E. Larsen l, Stewart S. Higgins ~, J. Thomas Raese 2 and John K. Fellman 3

~ Department of Horticulture and Landscape Architecture, Washington State University, Pullman, WA 99164 (U.S.A.)

2 USDA-ARS, Tree Fruit Research Laboratory, Wenatchee, WA 98801 (U.S.A.) 3Department of Plant, Soil and Entomological Sciences, University of ldaho, Moscow, ID 83843

(U.S.A.)

(Accepted for publication 26 July 1990)

ABSTRACT

Zeller, J.K., Larsen, F.E., Higgins, S.S., Raese, J.T. and Fellman, J.K., 1991. Rootstock effects on responses of potted 'Smoothee Golden Delicious' apple to soil-applied triazole growth inhibitors. II: Mineral nutrition and carbohydrate status. Scientia Hortic., 46: 75-88.

Emulsions containing 0 or 10 mg paclobutrazol, uniconazole or triapenthenol were applied to the soil of potted cultivar 'Smoothee Golden Delicious' apple (Malus domestica, Borkh. ) on M 9, M 7a or MM l 11 rootstocks (M =Malling, MM = Malling-Merton ). Significant differences in mineral con- centratic ns occurred, depending on rootstock, tissue and plant growth regulator (PGR), but few con- sistent trends emerged. PGR had no effect on carbohydrates in plants on MM 11 I. The carbohydrate concentrations in plants on M 7a were most sensitive to PGR, although differences in relative carbo- hydrate concentrations were also evident in trees on M 9.

Keyworcls: apple; carbohydrates; growth inhibitor; Malus domestica; minerals; paclobutrazol; tria- penthenol; uniconazole.

Abbreviations: M = Mailing; MM = Malling-Merton; PGR = plant growth regulator.

INTRODUCTION

A m o n g p l a n t g r o w t h r e g u l a t o r s ( P G R s ) for use o n app les , t r i a z o l e s h a v e

b e c o m e o f i n t e r e s t for t h e i r p o s s i b l e use i n t ree size c o n t r o l . T h e r e s p o n s e o f

*H/LA Paper No. 88-38, Project 1639, College of Agriculture and Home Economics Research Center, Pullman, WA 99164-6414, U.S.A.

0304-4238/91/$03.50 © 1991 - - Elsevier Science Publishers B.V.

76 J.K. ZELLER ET AL.

apple trees to these chemicals varies with scion and rootstock (Miller and Swietlik, 1986; Shearing and Palmer, 1986), but rootstock effects have not been well defined.

The effect of paclobutrazol on apple nutrition has been studied, often with conflicting results, which Davis and Steffens ( 1988 ) speculated may be due to different dosages, application timing, plant materials and environments. The carbohydrate response to triazole treatment has also been given atten- tion. Spur and shoot leaves of paclobutrazol-treated apple trees contained in- creased sorbitol and starch, while fructose, glucose and sucrose were generally unaffected when carbohydrates were expressed on a fresh weight basis (Stef- fens et al., 1985 ). Among carbohydrates in the wood, only glucose and starch increases were significant, and then only in early Spring (Steffens et al., 1985 ). Wang et al. ( 1985 ) observed that paclobutrazol-treated apple seedlings had a 55% increase in total carbohydrates. Leaves contained 41% of the total car- bohydrates, while roots had 44%. Sorbitol consistently increased in new stem tissue (70%), while starch increased primarily in roots and stems. Total car- bohydrates of paclobutrazol-treated roots were twice those of the control roots, indicating an alteration of assimilate partitioning. A pronounced shift in as- similates to the roots apparently caused increased root initiation and growth. In a related study (Wang et al., 1986b ), shoot growth ofpaclobutrazol-treated apples was not inhibited during the treatment year, but was reduced the sec- ond year. Non-structural carbohydrates increased in both years, indicating an effect independent of the growth effect.

Paclobutrazol altered the composition of apple tree cell wall polysacchar- ides (Wang et al., 1986a). Reducing sugars and starch generally increased in leaf and root tissue of apple trees treated with soil- and stem-applied paclo- butrazol (Wieland and Wample, 1985 ). The lowest paclobutrazol treatment caused the lowest starch and the highest reducing sugars, while the highest treatment caused the highest starch and lowest reducing sugars. Paclobutrazol and triapenthenol generally increased fructose, glucose and sucrose in apple stem tissue and fruit juice (Curry, 1988 ). Davis and Steffens (1988), citing Upadhyaya et al. ( 1986 ), speculated that increased starch produced in paclo- butrazol-treated plants may be due to decreased starch hydrolysis, which is the result of reduced amylase activity in treated plant tissues.

The research reported here compared nutrient levels in leaf and shoot tis- sues, and non-structural carbohydrates in leaves, shoots and roots of green- house-grown cultivar 'Smoothee Golden Delicious' apple on three rootstocks treated with soil applications of three triazole PGRs in order to explore the effect of rootstock and PGR on the response to triazole treatment.

M A T E R I A L S A N D M E T H O D S

Emulsions containing 0 or 10 mg paclobutrazol, uniconazole or triapen- thenol were applied to the soil of potted, greenhouse-grown 'Smoothee Golden

RESPONSES OF APPLE TO TRIAZOLE.II. 77

Delicious' apple trees on M 9, M 7a or MM 111 rootstocks (M=Mall ing, M = Malling-Merton). Tree growing, treatment methods and plot design were described previously (Zeller et al., 1991 ).

Mineral analysis

Leaf ( 10 leaves per sample) and shoot (five 2-cm segments per sample) tissues were rinsed with distilled water, blotted dry, lyophilized and ground in a Wiley Mill to pass a 40-mesh screen. The leaves were harvested at the completion of the 1986 experiment (20 october), during mid-day, from the mid-portion of each shoot; the shoot tissue was also from that same area. There was insufficient root tissue to analyze both mineral and carbohydrate contents. As root carbohydrate content was considered of greater interest, no mineral analysis of root tissue was performed.

Tot~Ll N was determined using 1-g samples placed into a LECO FP-228 N Determinator equipped with an autoloader. Nitrogen was analyzed by ther- mal conductivity and expressed as percent dry weight of tissue.

Ca, Mg, K, P, B, Zn and Fe contents were determined using a Beckman Spectraspan V Sequential Spectrometer (Beckman Instruments, Inc., Fuller- ton, CA ) equipped with a DC argon plasma emission source. Lyophilized leaf and shoot tissues, which had been prepared as above, were weighed into ce- ramic crucibles as 0.5-g samples, dry ashed at 500°C for 5 h, dissolved in 25 ml of an aqueous buffer solution containing 5 g 1-1 lithium ion plus 20% hy- drochloric acid, filtered through paper and then aspirated into the plasma (DeBolt, 1980 ). The resulting data were expressed as percentage of dry tissue weighl for Ca, K, Mg, N and P, while B, Zn and Fe were recorded as #g g- 1 of tissue.

Non-structural carbohydrate analysis

Non~-structural carbohydrates were determined at the conclusion of the ex- periment for leaves and shoot sections collected at mid-day from a mid-shoot location, and for the unsuberized new roots.

Tissue samples ( 100 mg each) were prepared as for mineral nutrit ion anal- ysis. These were extracted three times ( 3.5 + 3.5 + 3.5 ml) with 80% ethanol. Five milliliters of water and 5 ml of chloroform were added to the combined extracts of soluble carbohydrates. A vacuum evaporator at 40 ° C was used to dry the aqueous phase, and the residue was dissolved in 2 ml pyridine con- taining 30 mg ml-1 hydroxylamine hydrochloride and an internal standard of 5 rng ml-~ fl-phenyl-D-glucoside. The newly created oximes were derivi- tized to tri-methylsilyl ethers (Brobst and Lott, 1966) and analyzed with a temperature-programmed dual-column Hewlett-Packard 5830A gas chroma- tograph (Hewlett-Packard, Palo Alto, CA) equipped with flame ionization

78 J.K. ZELLER ET AL.

detectors and 1.6 mm X 1.8 m stainless steel columns packed with 3% OV-17 on 80/100-mesh Chromosorb.

A modification of a method developed by Oakley ( 1983 ) was used for starch determination. Two milliliters of sodium acetate buffer (0.1 M, pH 5.0) were added to the ethanol-insoluble residue of each original 100-mg sample and incubated at 100°C for 1 h. When cooled, 10 units of amyloglucosidase in acetate buffer were also added and again incubated at 55 °C for 16 h. A glu- cose oxidase assay was used to determine the amount of glucose resulting from the enzymatic degradation of starch. Each sample was diluted to contain < 200 /zg glucose ml -~, and to a 0.25-ml aliquot of this dilution were added 2.5 ml of a solution containing glucose oxidase, peroxidase and o-dianisidine dihy- drochloride (Sigma Chemical Co. Technical Bulletin No. 510, St. Louis, MO ). The samples were incubated at 25 °C for 1 h and absorbence was measured at 440 nm in a Hitachi 110A spectrophotometer (Hitachi Instruments Inc., Mountain View, CA). The amount of glucose in each aliquot was determined from a standard curve and starch was calculated by multiplying the amount of glucose by 0.9. Statistical analyses (see below) were conducted on leaf, stem and root contents of total non-structural carbohydrates (computed as the sum of sorbitol, fructose, glucose, inositol, sucrose, raffinose and starch), starch and non-reducing sugars (computed as the sum ofsorbitol and sucrose ).

Statist ical

Minerals. - Definite trends in the data were not evident from separate anal- yses of variance on each of the minerals assayed. Consequently, an agglom- erative classification scheme (sensu Pielou, 1984) was used to cluster the plants used in the experiment. The clustering was done by applying the aver- age linkage method of PROC CLUSTER (Statistical Analysis Systems Insti- tute Inc., 1982 ) to the raw mineral data collected from leaves and stems. The clusters that emerged from this procedure were little affected by whether the stem data were included. Our results, therefore, are based entirely on leaf samples. Results of the cluster analysis were used to point out which of the analysis of variance results (below) were likely to be the most important with respect to the treatments applied. In other words, where interaction terms were significant, the cluster analysis was relied upon to suggest which inter- actions were the most important to consider.

The effect of rootstock on the mineral content of the control plants was assessed by one-way analysis of variance.

For treated plants, we computed the difference in mineral content between each plant and the mean of the control plants according to the equation

x , j , = Y , j , - L . .

where Xak is the resultant difference for the kth tree which received the jth

RESPONSES OF APPLE TO TR1AZOLE.II. 79

PGR and was on the ith rootstock, Yu~ is the mineral of interest for the kth tree which received the jth PGR and was on the ith rootstock, and I7"i.. is the mean of the mineral of interest for the ith rootstock. This equation provides a negative result if the treated tree had a lower mineral content that its respec- tive control and a positive result otherwise. The treatment mean Xuk is termed AM below, for minerals.

The difference in mineral content between the treated and the control trees, i.e. Xu~, was analyzed as a 3 X 3 factorial experiment in a completely random- ized design, with rootstock (either MM 111, M 7a, or M 9) and PGR (either triapenthenol, paclobutrazol or uniconazole) as the factors.

Carbohydrates. - Multivariate analysis of variance (MANOVA) (Freund and Littell, 1981 ) was used to analyze simultaneously the responses of leaf, stem and root carbohydrates to rootstock and PGR. The experimental design was a 3 x4 factorial, with rootstock and PGR (including control plants) as the factors, in a completely randomized design. In instances where MANOVA indicated a significant interaction between PGR and rootstock, a subsequent MANOVA was conducted to assess the effect of PGR (including controls) within rootstocks. If significant differences among treatments were indicated by the,;e subsequent MANOVA, then a Fisher's Protected least significant dif- ference (LSD) test (Ott, 1984) was used for all possible pairwise compari- sons within rootstock.

Univariate analyses of variance (ANOVA) were also conducted, but we regarded the results of ANOVA to be of less relevance than the results of MANOVA since the latter more closely reflects the plant as an integration of leaves, stems and roots. For brevity, ANOVA results are presented only where they are of particular interest.

RESUETS

Leaf rninerals

Data from 36 trees were available for the cluster analysis, initially provid- ing 36 clusters each containing only one tree (only the last 12 clusters are shown in Fig. 1 ). Similarities among rootstocks did not emerge until there were only three clusters, at which point trees on M 9 which had no PGR treat- ment (the controls) or were treated with triapenthenol were grouped with trees on M 7a (Fig. 1 ). Furthermore, there was remarkably little ambiguity or overlap in the clustering of PGRs within rootstocks. Trees on MM 111 treated with uniconazole did not cluster together until there were only four clusters. Also, one tree on M 9 treated with triapenthenol was more similar to M 9 control trees than the other two M 9 trees which were treated with tria- penthenol, although M 9/control and M 9/triapenthenol had formed a clus-

8 0 J .K . Z E L L E R E T AL.

1

4

5 ¢0 _= o 6

"6 ~ 7

.O E " 8

Z

10

11

12 1_ U P P P C C C U T T T U U U U P P P T T T C C C T C C C T T U U U P P P

I MM 111 I M 7a I M 9 I

Rootstock/PGR t rea tment

Fig. 1. Dendrogram generated by agglomerative clustering of leaf samples of Ca, K, Mg, N, P, B, Fe and Zn taken from potted 'Smoothee Golden Delicious' apple trees on either MM 111, M 7a or M 9 rootstocks which were untreated (C) or treated with 10 mg per tree of paclobutrazol (P) , triapenthenol (T) or uniconazole (U) . Thirty-six trees were used in the experiment; the clustering procedure therefore began with 36 clusters. Only the final 12 clusters are preffented here.

TABLE1

Contents of selected leaf minerals of untreated, potted 'Smoothee Golden Delicious' apple trees on three rootstocks. Means (n = 3 ) within a column and followed by the same letter are not significantly different (P-< 0.05)

Rootstock Ca K Mg N P B Fe Zn (%) (%) (%) (%) (%) (/~g g- ~ ) (#g g-~) (/lgg - t )

M M l l l 1.37b 1.89a 0.55c 2.41b 0.15b 60a 129a 19b M7a 1.47h 1.66b 0.63h 2.60a 0.13c 41h 88b 18b M9 1.93a 1.78 ab 0.74a 2.55a 0.19a 36b 96b 26a

ter prior to the time when 12 clusters were distinguished (Fig. 1 ). Results o f the cluster analysis indicated quite clearly that the important interactions to be studied after analysis of variance were the effects o f PGR within rootstock.

Analysis of the mineral data indicated that untreated trees on MM 111 had higher B and Fe than trees on other rootstocks, but had lower N and Mg (Ta- ble 1 ). Control trees on M 7a did not have increased leaf mineral concentra-

RESPONSES OF APPLE TO TRIAZOLE.ll. 81

tions compared with other rootstocks, but had significantly less P. Untreated trees on M 9 had more Ca, Mg, P and Zn than other trees, but no element was lower :in trees on M 9 than on other rootstocks.

AM differed among minerals in its response to rootstock and PGR. Of the eight minerals assayed, the AM of B and Zn was not affected by an interaction between rootstock and PGR. The simple effects of rootstock and PGR were, however, significant for these two minerals (Table 2). Trees on MM 111 had the largest AM for B and the smallest for Zn. Trees on M 9 showed the smallest AM for B and the largest for Zn (Table 2 ). This clear distinction among root- stocks is consistent with the dendrogram (Fig. 1 ), which retained a distinc- tion among rootstocks until only three clusters were assigned. Among PGRs, triapenthenol had the greatest effect on B and the least on Zn, while no signif- icant differences emerged between paclobutrazol and uniconazole in their ef- fect on B or Zn (Table 2 ). The similarity of the effect of paclobutrazol and uniconazole was also noted in the dendrogram (Fig. 1 ), especially when ap- plied 1:o trees on M 9. The application of PGRs consistently increased leaf B and Zn.

The effect of PGR varied with rootstock, i.e. there was a significant root- stock:< PGR interaction for the AM of Ca, K, Mg, N, P and Fe (Table 3). For trees on MM 111, there was no significant effect of PGR on the d~ of Mg, P and Fe, although some of these AMs represented significant increases or losses of the respective mineral compared with the control trees (Table 3 ). Trees on M 7a showed no significant differences among PGRs for dMs of P and Fe. For trees on M 9, there was a significant effect of PGR on AMs for all minerals excepl: N (Table 3 ). Where significant effects of PGR within rootstocks were observed, the effects were consistent with the tree structure of the dendro- gram Jin all cases, except for Ca in leaves of trees on M 7a and M 9 (cf. Fig. 1, Table 3 ).

TABLE 2

Mean differences from untreated trees in leaf B and Zn of potted 'Smoothee Golden Delicious' apple trees on three rootstocks and treated with soil application of 10 mg per tree of one of the three plant growth :regulators (PGR), triapenthenol (T), paclobutrazol ( P ) or uniconazole (U). Rootstock and PGR showed no interactive effects. Means (n = 9 ) within a column and followed by the same letter are not significantly different (P-< 0.05 ). Positive values indicate an element content higher than in untreated controls. Asterisks indicate that a particular value is significantly different from the respec- tive control

Rootstc,ck B Zn PGR B Zn (/~g g- ~ ) (ugg -~ ) (ugg -~ ) (,ugg -~ )

MM l l l 21 a* 2c T 19a* 4b M 7a 13 b* 6 b* P 11 b 6 a* M9 2c 8a* U 6b 6a

82 J.K. ZELLER ET AL.

TABLE 3

Mean differences from untreated trees in leaf content of several minerals of potted 'Smoothee Golden Delicious' apple trees on three rootstocks and treated with soil applications of 10 mg per tree of one of the three plant growth regulators (PGR), triapenthenol (T), paclobutrazol (P) or uniconazole (U). Rootstock and PGR showed interactive effects. Means (n = 3) within a column and for PGR within rootstock, followed by the same letler, are not significantly different (P< 0.05 ). Positive val- ues indicate an element content higher than in untreated controls, with negative values indicating the reverse. Asterisks indicate that a particular value is significantly different from the respective control

Rootstock PGR Ca K Mg N P Fe (%) (%) (%) (%) (%) (/~g g-,)

MM 111 T -0 .42 b* 0.23a -0 .09 a* 0.15 a* 0.00a 1 a P -0 .19 a* -0 .04 b -0 .08 a* 0.03b 0.02 a* - 1 5 a * U -0 .17 a 0.00b -0 .03 a 0.04b 0.01a - 9 a

M 7a

M9

T 0.19 a* 0.16 b* 0.06 a* - 0.02 a 0.02 a* 22 a* P -0 .08 b 0.34 ab* -0 .06 b* 0.01 a 0.01 a* 27 a* U 0.00 ab 0.38 a* -0 .05 b -0 .23 b* 0.01 a 21 a*

T - 0 . 1 2 a 0.13a -0 .05 b 0.03a -0 .02 a 0a P -0 .28 ab* -0 .50 c* 0.06 a -0 .02 a -0 .04 b* - 2 0 b* U -0 .37 b -0 .26 b* 0.01 ab -0 .05 a -0 .06 b* 8a

Carbohydra te content

Total carbohydrates. - The effects of PGR on the concentrations of total car- bohydrates in leaves, stems and roots varied among rootstocks, i.e. MAN- OVA indicated a significant rootstock x PGR interaction (P< 0.0001 ). Sub- sequent analyses of the effects of PGR within rootstocks indicated that the patterns of total carbohydrate concentration among leaves, stems and roots for trees on MM 111 were not affected by PGR (Fig. 2). The response to PGR was substantial, however, among trees on M 7a. Not only did all PGRs cause significant shifts in the patterns of total carbohydrate concentration among leaves, stems and roots, but each PGR resulted in a distinct pattern (Fig. 2 ). Of particular interest among trees on M 7a are comparisons between control trees and trees treated with paclobutrazol, and between trees treated with paclobutrazol and uniconazole. Leaves on control trees did not differ in total carbohydrate concentration from leaves on trees treated with paclobu- trazol (P=0.3433), but differences between shoots and roots of these two treatments were both significant (P< 0.0001 and P = 0.0468 for shoots and roots, respectively) (Fig. 2). Roots of M 7a treated with paclobutrazol or uniconazole showed essentially identical concentrations of total carbohy- drates (P= 0.9846), whereas leaves and stems in these treatments showed a complete reversal in relative carbohydrate concentration (Fig. 2 ).

All three PGRs affected total carbohydrates among plants on M 9 roots,

RESPONSES OF APPLE TO TRIAZOLE.I I . 83

MM 11

M 7a

Control Tr iapenthenol Paclobutrazol Uniconazole

A 5.7 A

! 1.7 1.9 .6

Leaves ~--~ Stems ~ Roots 1017 10.0 512

5. 6 5.6

M9

4 6 ~ 2"2

3.2 7..5 7.0

Fig. 2. Kelative proportion of the total non-structural carbohydrate concentration in the leaves, stems and roots of potted 'Smoothee Golden Delicious' apple trees on either MM 111, M 7a or M 9 rootstocks which were untreated (controls) or were treated with 10 mg per tree of t r iapen- thenol, :paclobutrazol or uniconazole. Pie diagrams within a row and having the same letter adjacenl: to the upper left quadrant are not significantly different ( P < 0.05). Numerals at the periphery of each pie diagram represent the mean concentration (percentage of dry mass, n = 5 ) of total carbohydrate in the respective organs.

but there was no difference in the respective effects of triapenthenol and pa- clobutrazol (Fig. 2 ). Uniconazole doubled the leaf concentration of total non- structural carbohydrates compared with the other PGRs, and doubled the shoot t-oncentration compared with the control plants.

Non-reducing sugars. - A rootstock X PGR interaction was detected by MAN- OVA (P< 0.0001 ) in the response of non-reducing sugars to treatments. PGR within MM 111, however, had no effect on the concentration of these sugars in leaves, stems and roots (Fig. 3). When applied to trees on M 7a roots, triapenthenol and uniconazole resulted in altered patterns of non-reducing sugar concentrations (compared with control), but paclobutrazol was inef- fective (Fig. 3). As with total carbohydrates, non-reducing sugars in plants on M ,9 roots were significantly different from control plants when triapen- thenol and paclobutrazol were applied, and there was no significant differ- ence between these two PGRs. Uniconazole, however, was ineffective (Fig. 3).

84 J.K. ZELLER ET AL.

Control Triapenthenol Paclobutrazol Uniconazole

@@ MM 111 • 0.6 .1 .2

2.6 2.1

Leaves 4.6 ~ S t e m s ~ Roo ts 4-.9

1.6 2.1 2.0

5.0

11. 6.1 11.

0.3 M 9 0.5

~ 1.8 .1 .6

1.6 2.3

Fig. 3. Relative proportion of the non-reducing sugar concentration (sorbitol and sucrose) in the leaves, stems and roots of potted 'Smoothee Golden Delicious' apple trees on either MM 111, M 7a or M 9 rootstocks which were untreated (controls) or were treated with 10 mg per tree of triapenthenol, paclobutrazol or uniconazole. Pie diagrams within a row and having the same letter adjacent to the upper left quadrant are not significantly different ( P < 0.05 ). Nu- merals at the periphery of each pie diagram represent the mean concentration (percentage of dry mass, n = 5 ) of non-reducing sugars in the respective organs•

S t a r c h . - The pattern of starch concentration among leaves, stems and roots was also characterized by a significant roo t s tock×PGR interaction (P< 0.0127). An univariate analysis of leaf starch concentration from this same experiment failed to detect an effect of either rootstock, PGR (includ- ing control), or an interaction ( P = 0.0604, 0.3989 and 0.6040, respectively). This univariate result indicates, therefore, that treatment effects on starch concentrations in the stems and roots account for the interaction detected by the multivariate analysis. As with the other carbohydrate categories, PGR had no effect on starch concentrations in plants on MM 111 rootstocks (Fig. 4). Among trees on M 7a, all three PGRs affected starch concentrations simi- larly, but only paclobutrazol and uniconazole significantly altered patterns of starch concentrations compared with controls (Fig. 4). PGR-treated trees on M 9 roots had different starch concentration patterns than control trees, but there were no differences among PGRs in their effect (Fig. 4).

RESPONSES O F A P P L E T O T R I A Z O L E . ] I . 8 5

Control Triapenthenol Paclobutrazol Uniconazole

M M 1 1 1 0.6 .7 5.6 .7 5.2 .7

4.3 .3 ,

Leoves ~ Stems ~ Roots 1.6

0 o 2. 4.5 4. .4 .5 .6 3.6

M 7G

1.1

1.3 2.1

6 B 6 1.5 " .0 0 . 7

4.3 5.6 " 4,7

M 9 .s ~.7 .8

1,1

Fig. 4. Relative proportion of the starch concentration in the leaves, stems and roots of potted 'Smoothee Golden Delicious' apple trees on either MM I 1 l, M 7a or M 9 rootstocks which were untreated (controls) or were treated with l0 mg per tree of triapenthenol, paclobutrazol or uniconazole. Pie diagrams within a row and having the same letter adjacent to the upper left quadrant are not significantly different (P_< 0.05 ). Numerals at the periphery of each pie dia- gram represent the mean concentration (percentage of dry mass, n = 5) of starch in the respec- tive organs.

DISCUSSION

Mineral content

B and Zn in the leaves of 'Smoothee Golden Delicious' showed opposite responses to both rootstock and PGR. Therefore, we may conclude that the generally greater growth response of plants on MM 111 to PGR (Zeller et al., 1991 ) was correlated with an increase in leaf B. Also, the generally greater marginal effectiveness for growth reduction of triapenthenol (Zeller et al., 1991 ) was correlated with an increased leaf B content in those plants treated with this chemical (Table 2). Further broad generalizations concerning leaf mineral content are difficult because for all other minerals the effect of PGR was dependent upon rootstock. When trees were on MM 111 roots, unicona- zole did not affect any of the measured minerals, but on M 9, triapenthenol was generally ineffective, increasing only leaf B (Tables 2 and 3 ). K was af- fected by uniconazole on both M 7a and M 9 rootstocks, but the effect was to increase K on M 7a and decrease it on M 9. Of the 27 values in Tables 2 and

86 J.K. ZELLER ETAL.

3 that were significantly different from control trees (represented by aster- isks), 13 (approximately one-half) were decreases in the respective mineral content. Furthermore, all columns in Table 3, regardless of which mineral they represent, include both increases and decreases in leaf mineral content.

The few obvious and consistent trends in leaf mineral content suggested either that the effects of PGR were not mediated directly by mineral nutrition or that we need to take a more mechanistic approach to the study of the ef- fects of PGRs on plant nutrition. A formal test regarding the former sugges- tion would require that our experiment be repeated on a series of plants that initially varied in their nutrient status. Any such study should, however, also at least monitor gibberellic acid and, possibly, auxin levels in the plant.

A comparison of Fig. 1 with Tables 2 and 3 may help to infer which min- erals may provide the most fruitful lines of research concerning the mechan- istic influence of PGR on mineral status. For plants on MM 111, those treated with paclobutrazol clustered very early with their respective controls (Fig. 1 ), suggesting that the overall effect of paclobutrazol on leaf mineral content was minimal. As K, N and B were the only minerals that did not differ from con- trol leaves, these elements may have been most responsible for the clustering noted for MM 111 in Fig. 1. On M 7a and M 9, paclobutrazol-treated plants did not cluster with their respective controls until agglomeration had pro- ceeded to relatively few clusters. Of the minerals that were consistent with the clustering pattern in leaves on MM 111, K was consistent with the clustering pattern on the other rootstocks, i.e. paclobutrazol-treated leaf K was signifi- cantly different from control plants on both M 7a and on M 9. Furthermore, the difference is large relative to the other macroelements (Table 3). It ap- peared that leaf K may be an important mineral in paclobutrazol's mode of action.

Carbohydrate s tatus

Total carbohydrates. - The concentrations of non-structural carbohydrates in the leaves, stems and roots of 'Smoothee Golden Delicious' grafted to MM 111 were not affected by the application of 10 mg PGR. Therefore, if non- structural carbohydrates were associated with the reduced growth in response to increasing PGR in plants on MM 111 (Zeller et al., 1991 ), then the rela- tionship must be a function of shifts in the relative sizes of plant parts and differences in the quantity, not concentration, of these carbohydrates in the respective organs.

PGR treatment appeared to reduce the concentration of carbohydrate in the roots o fM 7a and M 9, but this appears to be about the only generalization possible with respect to total carbohydrates in trees on M 7a and on M 9 (Fig. 2 ). Trees on M 7a appeared to be more sensitive to the respective PGRs than trees on M 9, and this agrees with the generalization by Zeller et al. ( 1991 )

RESPONSES OF APPLE TO TRIAZOLE.II. 87

that the more vigorous rootstocks were more sensitive to triazole growth in- hibitors (i.e. had a greater percentage reduction in shoot growth).

N o n - r e d u c i n g s u g a r s . - As with total carbohydrates, PGR had no effect on the concentrations of non-reducing sugars (sorbitol and sucrose ) in trees on MM 111 roots. Also, for trees on M 9, the effect of paclobutrazol was indistin- guishable from that of triapenthenol. Once again, M 7a was the most sensitive rootstock to uniconazole treatment.

S t a r c h . - Starch was the carbohydrate least affected by PGR. In fact, leaf starch was unaffected by rootstock or PGR. Root starch was reduced by PGR in both M 7a and M 9, and stem starch was increased, but the three PGRs tested were equally effective in eliciting this response.

In conclusion, plant carbohydrate status may be most easily manipulated in plants on M 7a roots and to a lesser degree in plants on M 9 roots. Also, plant starch concentration appeared to be a poorer indicator of the effect of PGR on non-structural carbohydrates than were either total carbohydrates or non-reducing sugars. However, it is still unclear what relationship, if any, these patterns of non-structural carbohydrate concentrations had with the growth- reducing effects of these PGRs. Future studies along these lines would do well to assess the effects of the PGRs on plant hormones, particularly gibberellic acid.

ACKNOWLEDGMENTS

Trees for this study were supplied by Hilltop West Nursery, Ephrata, WA, U.S.A. We thank D. Staiff, R. Eisenreich, T. McCamant and W. Loescher for their waluable technical assistance, and R. Alldredge for his statistical advice.

REFERENCES

Brobst, K.M. and Lott, C.E., 1966. Determination of some components in corn syrup by gas- liquid chromatography of trimethylsilyl derivatives. Cereal Chem., 43: 35-42.

Curry, E.A., 1988. Chemical control of vegetative growth of deciduous fruit trees with paclo- butrazol and RSW0411. HortScience, 23: 470-473.

Davis, T.D. and Steffens, G.L., 1988. Triazole plant growth regulators. Hortic. Rev., 10: 63- 105.

DeBolt, D.C., 1980. Multielement emission spectroscopic analysis of plant tissue using DC ar- gon plasma source. J. Assoc. Off. Anal. Chem., 63: 802-805.

Freund, R.J. and Littell, R.C., 1981. SAS for linear models A guide to the ANOVA and GLM procedures. SAS Institute Inc., Cary, NC, 231 pp.

Miller, S..S. and Swietlik, D., 1986. Growth and fruiting response of deciduous fruit trees treated with paclobutrazol. Acta Hortic., 179: 563-566.

8 8 J.K. ZELLER ET AL.

Oakley, E.T., 1983. Enzymatic determination of starch in fresh green lyophilized green, and cured tobacco. J. Agric. Food Chem., 31: 902-905.

Ott, L., 1984. An Introduction to Statistical Methods and Data Analysis. Duxbury Press, Bos- ton, MA, 775 pp.

Pielou, E.C., 1984. The Interpretation of Ecological Data A Primer on Classification and Ordi- nation. Wiley, New York, 263 pp.

Shearing, S.J. and Palmer, B., 1986. International survey of paclobutrazol results on apples. HortScience, 21: 779 (Abstract).

Statistical Analysis Systems Institute Inc., 1982. SAS user's guide: Statistics. SAS Institute Inc., Cary, NC, 584 pp.

Steffens, G.L., Wang, S.Y., Faust, M. and Byun, J.K., 1985. Growth, carbohydrate, and mineral element status of shoot and spur leaves and fruit of 'Spartan ' apple trees treated with paclo- butrazol. J. Am. Soc. Hortic. Sci., 110: 850-855.

Upadhyaya, A., Davis, T.D. and Sankhla, N., 1986. Some biochemical changes associated with paclobutrazol-induced adventitious root formation on bean hypocotyl cuttings. Ann. Bot., 57: 309-315.

Wang, S.Y., Byun, J.K. and Steffens, G.L., 1985. Controlling plant growth via the gibberellin biosynthesis system - II. Biochemical and physiological alterations in apple seedlings. Phys- iol. Plant., 63: 169-175.

Wang, S.Y., Steffens, G.L. and Faust, M., 1986a. Effect of paclobutrazol on cell wall polysac- charide composition of the apple tree. Phytochemistry, 25: 2493-2496.

Wang, S.Y., Steffens, G.L. and Faust, M., 1986b. Effect of paclobutrazol on accumulation of carbohydrates in apple wood. HortScience, 21: 1419-1421.

Wieland, W.F. and Wample, R.L., 1985. Effects ofpaciobutrazol on growth, photosynthesis and .carbohydrate content of 'Delicious ' apples. Scientia Hortic., 26: 139-147.

Zeller, J.K., Larsen, F.E., Higgins, S.S. and Curry, E.A., 1991. Rootstock effects on response of potted 'Smoothee Golden Delicious' apple to soil-applied triazole growth inhibitors. I: Shoot and root growth. Scientia Hortic., 46:61-74.