Defining Phosphorus Requirements for Pinot noir Grapevines

Defining Phosphorus Requirements for Pinot noir Grapevines R. Paul Schreiner 1 * and James Osborne 2 Abstract: A study to examine the phosphorus (P) requirements of Pinot noir was carried out using a microplot vineyard with carefully controlled P inputs. Pinot noir grafted onto 101-14 rootstock was exposed to four levels of P supply delivered via fertigation beginning in their fourth growing season. Vine nutrient status, productivity, and must chemistry were studied over four years (2012 to 2015), and fermentation dynamics were evaluated over three years (2012 to 2014). P supply primarily influenced vine productivity by reducing leaf area at veraison and by reducing yield, which occurred after three years in vines that received no P fertilizer. Flowering and fruit set were not altered by low P status. P supply had the most significant impact on must P levels, where the two lowest P supply treatments had reduced must P within the first year that P was altered. However, must P concentrations as low as 32 mg P/L did not affect the time needed for yeast to complete alcoholic fermentation. These findings suggest that limiting P reduces Pinot noir canopy size and yield before it alters flowering parameters or reduces must P concentrations sufficiently to alter fermentation. A P concentration of 1.0 g P/kg dry weight (DW) in leaf blades at veraison is proposed as the critical level below which growth and yield of Pinot noir are reduced when vines are cropped at levels typical for premium wine production in the region. Growers should closely monitor vine P status when leaf blade P at veraison approaches 1.2 g P/kg DW in western Oregon Pinot noir vineyards to account for sampling and laboratory error. Key words: growth, leaf area, leaf phosphorus, must phosphorus, Vitis vinifera, yield Phosphorus (P) limitation in vineyards has historically been uncommon in many grapegrowing regions worldwide (Cook 1966), and reports of vine growth or yield increases in response to P fertilization are equally rare or of a small magnitude (Conradie and Saayman 1989a, 1989b, Klein et al. 2000, Schreiner at al. 2013). Grapevines are quite adept at obtaining P from soil, in large part because they are exceptional hosts for arbuscular mycorrhizal fungi harboring very high levels of these symbiotic fungi in their roots (Schreiner and Scagel 2016). However, there are cases where P limitation brought about by very low soil P availability and/or water limitation may affect grapevine productivity, including the hillside vineyards of western Oregon (Schreiner et al. 2006). In addition, P limitation may become more common in western Oregon vineyards as more acreage is planted on shallower 1 USDA-ARS-Horticultural Crops Research Unit, 3420 NW Orchard Avenue, Corvallis, OR 97330; and 2 Department of Food Science, 100 Wiegand Hall, Oregon State University, Corvallis, OR. *Corresponding author (Paul.Schreiner@ars.usda.gov; tel: 541-738-4084; fax: 541-738-4025) The authors gratefully acknowledge Mathew Scott, Suean Ott, and Keira Newell for technical assistance. They also thank Duarte Nursery, Inc. (Hughson, CA) for providing certified grapevines. This work was funded, in part, by the Oregon Wine Board, the Oregon Wine Research Institute, and USDA-ARS CRIS 2072-21000-048-00D. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. Supplemental data is freely available with the online version of this article at www.ajevonline.org. Manuscript submitted Jan 2018, revised Apr 2018, accepted May 2018 doi: 10.5344/ajev.2018.18016 hillside soils and on low-vigor rootstocks. P deficiency was observed in Pinot noir grafted onto the two most common rootstocks used in western Oregon (3309C and 101-14 Mgt) in a field microplot trial using a red hill soil, but not on other rootstocks or own-rooted vines (Schreiner et al. 2012). When P is limited in vineyards, yield is generally thought to be more strongly reduced than vine vegetative growth, owing to the greater sensitivity of flowering and fruit set to P status (Tulloch and Harris 1970, Skinner et al. 1988, Skinner and Matthews 1989). Low fruit set is common in western Oregon vineyards, but it is unclear if P limitation is a factor beyond the weather conditions around the time of bloom. Results from a prior study examining P requirements of Pinot noir led to the suggestion that must P concentrations might be a good metric to better target P status in grapevines, as this was more sensitive to varying P status than was growth or yield (Schreiner et al. 2013). A value for must P of 100 mg P/L was proposed as a starting point, assuming that this much P was needed for optimal fermentations based on work examining P removal from musts (Archer and Castor 1956) and general guidelines for yeast growth (Jones and Gadd 1990, Bisson 1999). However, we found no reports that clearly define yeast P requirements in grape musts from red wine cultivars. Therefore, minimum P requirements based on yeast needs to complete healthy alcoholic fermentations (analogous to minimum yeast assimilable nitrogen [YAN] requirements) are not known for Pinot noir. The goals of this research were to evaluate vegetative and reproductive responses to varying levels of P supply, and to understand how fruit chemistry and must fermentation are altered by P status. Grafted Pinot noir vines were grown over four years in a microplot vineyard where P supply was 351

352 Schreiner and Osborne carefully controlled. To ensure that P limitation would be obtained (Schreiner et al. 2013), the lowest P supply treatment received no P additions to the soil. An additional goal was to define leaf blade and petiole P tissue guidelines at bloom and veraison for grafted Pinot noir vines carrying a typical crop level for the region. Materials and Methods Microplot vineyard and experimental design. This study was conducted over four consecutive growing seasons (2012 to 2015) in a microplot (pot-in-pot) vineyard where Pinot noir was grown with varying levels of P supply. The microplot vineyard and vine management protocols were described previously (Schreiner and Scagel 2017, Schreiner et al. 2018). Briefly, grafted Pinot noir grapevines (Vitis vinifera L. Pommard clone, FPS 91, on 101-14 Mgt rootstock) were established and grown for three years (2009 to 2011) in a sand:soil mixture that contained 9 mg/kg of available P (Bray 1) prior to altering P supply. Vines were given a complete nutrient solution (halfstrength Hoagland s solution with 0.50 mm P; Hoagland and Arnon 1950) supplied via fertigation for the first three years. In year four and thereafter, the concentration of P was supplied at one of four rates (designated as Control [100% P], 50% P, 20% P, and no P), where the total concentration of P in the Control treatment during fertigation events was 0.50 mm and the lower rates were 0.25, 0.10, and 0.0 mm total P. All other nutrients (N, K, Ca, Mg, S, Fe, Mn, B, Zn, Cu, Mo) were held constant. Irrigation inputs were adjusted in all treatments to maintain similar soil and vine water status among different nutrient treatments as described (Schreiner et al. 2018). Fruit exposure to sunlight was also controlled by applying different levels of leaf removal in the cluster zone (mainly on the east side of the canopy) so that all treatments had similar light levels in the fruit zone throughout the day based on ceptometer measurements (AccuPAR Model LP-80, Decagon). Each treatment was replicated four times in a randomized complete block design, and each replicate plot comprised five continuous vines. Vines were cane-pruned and trained on a single Guyot system with vertical shoot-positioning. Fruit clusters were thinned approximately two weeks after fruit set each year to one cluster per shoot, including the renewal shoot (spur), by retaining either the basal cluster or the second cluster. Fungicides were used to manage powdery mildew (Erysiphe necator [Schw. Burr.]) and bunch rot (Botrytis cinerea L.), as per standard practices in the region. Vine nutrient status. Vine leaf blades and petioles were collected to determine nutrient status at 50% bloom and at 50% veraison each year. Ten leaves per plot were sampled and combined from count shoots at both bloom and veraison between 0900 and 1100 hr. Leaves opposite clusters were collected at bloom, and paired leaf samples comprising a leaf opposite a cluster and a recently expanded leaf were collected at veraison. Leaf blades and petioles were separated, rinsed in distilled water, dried at 65 C for 48 to 72 hr (Shel Lab FX 28-2, Sheldon Manufacturing Inc.), and ground to pass through a 425-μm-sieve. Concentrations of P and other nutrients (K, Ca, Mg, S, Fe, Mn, B, Zn, Cu) were measured by ICP-OES (inductively coupled plasma-optical emission spectrometry; Perkin Elmer Optima 3000DV) after microwave digestion in HNO 3 (Jones and Case 1990). N was determined via combustion analysis (Leco, Inc.). Reference standard apple (Malus domestica L.) leaves (no. 151, National Institute of Standards and Technology) were included in each set of samples to ensure instrument and digestion procedures were accurate. Leaf blade and petiole concentrations are expressed on tissue dry weight (DW) basis. Vine vegetative growth and photosynthesis. Shoot length and leaf area per vine was measured at bloom, and leaf area was measured at veraison in each year by first obtaining the primary shoot length and the length of all lateral shoots for all shoots on the middle three vines per plot. The area of leaves on main shoots and lateral shoots was then determined on 20 random shoots per treatment (80 total, ensuring that both larger and smaller shoots were included from each treatment) by comparing leaves to a series of concentric circles with known area as described (Schreiner and Scagel 2017). Leaf area per vine was calculated from the relationships between leaf area and shoot length for main shoots and for lateral shoots, and summed for all shoots per vine. Dormant season pruning mass (fresh weight of one-year-old canes) from the three middle vines per replicate was determined in the winter by weighing the count shoots from the previous season. Leaf gas exchange was measured using a portable infrared gas analyzer system (LiCor 6400, LiCor Inc.) on a single leaf per plot. Fully exposed leaves (PAR >1800 μmol/m 2 s) on a main shoot in the lower or middle canopy were measured at bloom and veraison, respectively. Measurements of gas exchange were made at various times during the day, but data collected within one hour of solar noon (1300 hr) are included here. Vine reproductive growth and yield parameters. Flowers and fruit set were determined by placing fine mesh fabric bags on two random clusters per plot prior to the onset of flowering. The bags were carefully removed after fruit set, ensuring that all flower caps were collected by inserting a small tray under each bag, and the total number of flower caps was counted. Each cluster used for this purpose was tagged to later sample just prior to harvest to count the final number of berries and calculate fruit set. The date of fruit harvest in each year was based on a random sampling of berries from all plots (three berries per plot), when berries reached 22 to 24 Brix. However, in 2013, high rainfall just prior to fruit maturity decreased berry soluble solids below 20 Brix, and fruit was harvested eventually at ~21 Brix. All plots were harvested on the same day each year. Fruit clusters were removed from the three middle vines per plot, counted, and weighed, to determine yield and average cluster weights. A subsample of five randomly selected clusters from each plot was transported back to the lab to determine the number of berries per cluster and average berry weight. Must chemistry and alcoholic fermentation. The fivecluster subsample from each plot was juiced using a stainless steel hand-crank press to obtain a yield equivalent to 625 ml must/kg fresh weight of clusters, using at least two pressings that were combined for analysis. Fruit maturity indices

Phosphorus in Pinot noir 353 (soluble solids, ph, and titratable acidity) were determined as described previously (Schreiner et al. 2013). YAN concentration in must was determined by summing free amino acid-n (FAA-N) obtained by the o-phthaldialdehyde (OPA) colorimetric assay (Dukes and Butzke 1998) and ammonium- N, by an enzymatic assay (Sigma ammonia assay kit; Sigma Chemical Co.). Must P, K, Ca, Mg, S, Fe, Mn, B, Zn, and Cu concentrations were measured by ICP-OES after microwave digestion in HNO 3 as per leaf blades and petioles. In 2012 to 2014, all remaining fruit per plot was combined, stored overnight at 4 C, and destemmed the next day. Field replicates were processed separately, and 3 kg of each replicate was placed in 4 L microfermenters as described (Sampaio et al. 2007). Fermenters were placed in a temperature-controlled room set at 27 C, warmed to room temperature, and inoculated with Saccharomyces cerevisiae RC212 (Lallemand) at ~10 6 cfu/ml. Fermentations were conducted with a submerged cap (Sampaio et al. 2007), and soluble solids were monitored daily using an Anton-Paar DMA 35N density meter. After all fermentations reached dryness (<0.5 g/l reducing sugar as measured by Clinitest, Bayer), they were pressed using a modified basket press with an applied constant pressure of 0.1 MPa for 5 min. Fifty mg/l SO 2 (as potassium metabisulfite) was added to the wines before they were cold-settled at 4 C for five days. Wines were then racked and an addition of SO 2 was made to achieve 25 to 30 mg/l free SO 2 prior to being bottled in 375 ml screwcapped (Stelvin, Amcor) wine bottles, and stored at 13 C. Statistical analysis. Data derived from multiple observations per plot (i.e., shoot length) were first averaged per replicate plot prior to analysis. Factorial analysis of variance (ANOVA) was used to examine how specific vine and fruit variables were altered by P supply and year, accounting for the block effect in the models. Variance assumptions were tested using Cochran s test, and residuals were examined to ensure normality. Must P concentration data were log-transformed and veraison petiole P concentration data were inverse-transformed prior to ANOVA to satisfy variance assumptions. Means were compared using Tukey s post-hoc test at 95% confidence. For simplicity, the means and standard errors of the mean are reported in all figures. Data from the season prior to altering P supply (2011) is shown as a reference in figures, although this data was not included in the analysis. Results The weather from 2012 to 2014 varied as described previously (see Schreiner et al. 2018): 2012 was the coolest year, 2014 was the warmest, and 2013 was intermediate between these years. High rainfall also occurred just before harvest in 2013. The 2015 growing season was similar to 2014 until the end of June, but was warmer in July, which led to the earliest date for veraison (color change) among all four years (Supplemental Table 1). Harvest in 2015 was only two days earlier than in 2014, since the average air temperature was lower between veraison and harvest in 2015 than in 2014. Volumetric soil water content (θ v ) was altered by P supply on several days in each year, but the differences were minor (<2.0%) and corrected the next day by adjusting irrigation rates in different treatments (data not shown). An analysis of θ v in each year using data from three growth periods (budbreak to bloom, bloom to veraison, and veraison to harvest) and the four P supply treatments as factors indicated that P supply did not alter θ v in any year, nor did it interact with growth period (Supplemental Table 2). Therefore, the small differences in soil moisture that had occurred on a few days did not lead to consistent effects across the season or within a developmental growth period in any year. Solar exposure of fruit clusters based on ceptometer measurements was not affected by P supply in any year (data not shown). P supply and year altered P status of Pinot noir in a manner that was largely consistent with expectations (Figure 1). All measures of P status were influenced by the interaction between year and P supply, as leaf blade and petiole P concentrations among the different P levels did not respond the same each year. For example, bloom leaf blade P in all three low-p treatments was reduced relative to the Control (100% P) treatment to a similar concentration in 2012, and declined further to a similar level in 2013 and 2014. However, by 2015, bloom leaf blade P concentrations were lower in the no-p and 20% P vines than in the 50% P vines, which were lower than the Control vines. Bloom petiole P concentrations did not respond the same as leaf blades, as only the no-p vines were lower than the Control vines in 2012, but all three low-p treatments were lower than the Controls in subsequent years, with the no-p vines having less petiole P than the 50% P vines only in 2014. Veraison leaf blade P showed the most distinct differences among all of the P supply treatments after 2012: the two lowest-p treatments (no P and 20% P) had lower leaf P than the 50% P vines in 2013, which were also lower than the Control vines that year. All four P supply treatments differed in 2014 and 2015. Petiole P concentrations at veraison showed the largest relative decrease in the low-p treatments compared to the Control, but the two lowest-p supply treatments did not differ in any year. Petiole P concentration data at veraison also required a transformation to satisfy variance assumptions for ANOVA, primarily due to the large variation in the Control (100% P) vines. Vegetative growth parameters examined here were reduced only in those vines receiving no P for three or four years (Figure 2). Bloom shoot length was altered by the interaction between P supply and year because the no-p vines differed from the Control only in 2015. Compared to the Controls, shoot length at bloom was reduced by 19% in the no-p vines in 2015. Leaf area at bloom was reduced only in the no-p vines (by 8%) as a main effect across all years, but this was driven mostly by lower values in 2014 and 2015. Leaf area at veraison was lower in the no-p treatment vines in 2014 and 2015, but not earlier, leading to an interaction between P supply and year. Veraison leaf area was reduced in the no-p vines by 23% in 2014, and by 28% in 2015, compared to the Control vines. Dormant season pruning mass was reduced in 2015 only in the no P vines by 37%, compared to the Control. Photosynthesis measured on fully exposed leaves was not altered by P supply in any year at bloom or at veraison (data not shown). Leaf symptoms of P

354 Schreiner and Osborne deficiency were first observed in year 2 (2013) at the time of fruit maturity, and this was repeated in 2014 and 2015. Symptoms first occurred in the no-p vines, but were also found on some individual vines in the 20% P treatment replicates in 2014 and 2015. Careful scouting of vines throughout the growing season in all years revealed that leaf symptoms appeared in the days just prior to harvest in 2013 and 2014, but about one week prior to harvest in 2015. The symptoms included leaf reddening, particularly on more exposed leaves (or even more exposed portions of single leaves), and symptoms were more common on older and middle-canopy leaves. The symptoms observed here differed somewhat from classic P deficiency symptoms in that the red color was not initially expressed at the leaf margins between larger veins, but rather anywhere on the leaf surface that was exposed to high solar radiation. A separate analysis of symptomatic and non-symptomatic leaves from vines within the same replicate collected at fruit maturity in 2013 and 2014 showed that symptoms occurred when leaf blade P was below 0.80 g P/kg DW, and petiole was below 0.40 g P/kg DW. This equated to leaf P concentrations at veraison equal to or below 1.0 g P/kg DW. Certain vine reproductive measures were reduced by low P supply, but this occurred only in the no-p vines as well (Figure 3). Fruitfulness (cluster number per shoot) was altered by P supply but not by year, although none of the low P treatments differed from the Control. Rather, the intermediate P supply vines at 50% P and 20% P had more clusters per shoot than the no-p vines, but no treatment differed from the Control vines. The number of flowers per inflorescence and the proportion of flowers that set fruit were not altered by P supply (data not shown). Yield was eventually reduced under low P supply, but this was only apparent in the no-p vines in 2014 and 2015, after three or four years with no added P. The relative impact of low P on yield was similar to its effect on veraison leaf area. Yield was reduced by 23% in 2014 and by 35% in 2015 in the no-p vines, compared to the Controls. The lower yield of no-p vines was driven more by lower berry weight than by berry number, since berry number per cluster was ~9% lower (as a main effect across years) in no-p vines, while berry weight was 20% lower in the no-p vines in 2015, compared to the Control. In addition, berry weight was reduced in the 20% P vines in 2015, but berry number per cluster was not. P supply did not alter must chemistry dramatically, aside from the expected impact on must P concentration (Figure 4). Soluble solids at harvest were not altered by P supply, Figure 1 Interactive effect of year and phosphorus (P) supply on leaf blade and petiole P concentrations at bloom and veraison in Pinot noir grapevines grown in microplots from 2012 to 2015. The Control treatment (100% P) received 0.50 mm P during fertigation events. Data from 2011, before P was manipulated, are shown as a reference. Letters near each symbol designate significant mean groups based on Tukey s honest significant difference at 95% confidence. Data are means and standard error of the mean for each plot (n = 4). Petiole P data at veraison were inverse-transformed prior to analysis of variance. DW: dry weight.

Phosphorus in Pinot noir 355 although differences by year were apparent, particularly in 2013 when heavy rainfall occurred just prior to harvest. Must ph in the no-p treatment was ~0.1 units greater than in the Control as a main effect across all years, while titratable acids were lower in both the 20% P and no-p musts, compared to the Control as a main effect across years. Must P concentrations were greatly reduced by lower P supply, such that P was reduced below 40 mg P/L in the no-p vines in 2013 through 2015, while the Control ranged from 115 to 135 mg P/L in those years. The juice P concentrations closely matched the response of leaf blade P concentrations at veraison. Decreasing must P concentration did not impact the rate of alcoholic fermentation. All fermentations were complete within five days after inoculation in each year (data not shown), despite P ranging from as low as 32 mg P/L in the 2015 no-p vines, to as high as 135 mg P/L in the Controls. P supply altered the status of other mineral nutrients in leaf blades, petioles, or musts, but most of these changes were not of great magnitude. For example, leaf blades and petioles at veraison had increased concentrations of potassium (K) and iron (Fe) and decreased calcium (Ca) and magnesium (Mg) by ~10 to 20% as P supply decreased (Supplemental Table 3 shows data for leaf blades). The concentration of boron (B) increased in leaf blades (Supplemental Table 3), but decreased in petioles (data not shown) as P supply decreased. There was a higher concentration of sulfur (S) and manganese (Mn) in leaf blades in the 50% P vines than in the no-p vines. While these changes were statistically significant as a main effect across all years or within specific years, nutrient concentrations were within a healthy range for all treatments, and none approached known deficient or toxic concentration values (Robinson 2005, Schreiner et al. 2018). In musts, Mg, S, Mn, B, and YAN were altered by P supply (Supplemental Table 4). However, only Mg and Mn were altered by P supply in musts in the same manner as observed in leaf blades and in petioles. Discussion The salient finding from this study was that altering vine P status in Pinot noir had the largest and most immediate impact on fruit P concentrations, reducing must P levels by 40% within the first year and by nearly 80% after four years in the no-p vines. Even those vines receiving 50% P had significantly less must P concentrations than the Control vines by the second year after altering P supply. By years three and Figure 2 Effect of year and phosphorus (P) supply on vegetative parameters of Pinot noir grapevines grown in microplots from 2012 to 2015. Interactive effect of year and P supply (scatter charts) on bloom shoot length, veraison leaf area, and pruning mass (n = 4), and main effect of P supply (bar chart) on bloom leaf area (n = 16). A * to the right of a symbol in interactive plots indicates those P treatments that differed from the Control (100% P) in each year based on Tukey s honest significant difference (HSD) at 95% confidence. Data from 2011, before P was manipulated, are shown as a reference in interactive plots. Letters above means for bloom leaf area indicate treatment differences based on Tukey s HSD at 95% confidence. Data are means and standard error of the mean for each plot. FW: fresh weight.

356 Schreiner and Osborne four, must P in 50% P vines was reduced to about half that of the Controls. A large impact of vine P supply on must P concentration has been reported previously (Skinner and Matthews 1989, Schreiner et al. 2013), leading to the suggestion that must (or juice) P concentration is a reliable indicator of grapevine P status. The impact of P supply on vine growth or fruit yield did not occur until after three years in those vines that received no P, and the magnitude of these effects was smaller than the impact on must P concentration. Pinot noir vines that received no P had a smaller canopy (leaf area) at veraison in years three and four after treatments began, while bloom shoot lengths and dormant season pruning mass were not reduced in the no-p vines until year four (2015). Yield was reduced also in years three and four only in vines receiving no P, similar to the impact on veraison leaf area. The reduction in leaf area and yield that occurred in 2014 and 2015 was associated with veraison leaf blade P concentrations below 1.0 g P/kg DW. Even though must P levels were altered most by P status, the rate of alcoholic fermentation by S. cerevisiae RC212 was not influenced by must P concentrations as low as 32 mg P/L. These findings indicate that P limitation on yeast alcoholic fermentation in red wines will rarely be a significant concern for winemaking. Previously, we suggested that must P requirements for yeast may be ~100 mg P/L, and that leaf blade and petiole P test guidelines could be based on achieving this level of must P (Schreiner et al. 2013). The value of 100 mg P/L was based on the quantities of P removed from grape musts (Archer and Caster 1956), and on optimum P levels, suggested in a review on yeast nutrition (Jones and Gadd 1990). Published studies of P requirements for fermentation of winegrape musts are unknown, but our data are supported by a study in Sake that showed that yeast (S. cerevisiae) growth and alcohol production were maximal at 31 mg P/L (Fujitani 1965). However, the rate of yeast growth was only slightly reduced at P levels as low as 7 mg P/L in that study. In another study employing defined media, cellular growth of S. cerevisiae was only slightly lower than that in a luxury P medium (341 mg P/L) when the P concentration was 6.8 mg P/L, but growth was more sharply reduced at 3.4 mg P/L (Toh-e et al. 1973). Therefore, minimum yeast P requirements to complete fermentation are well below 30 mg P/L. Since fermentations were not affected by must P concentrations achieved in this study, P management and tissue P guidelines for Pinot noir Figure 3 Effect of year and phosphorus (P) supply on reproductive parameters of Pinot noir grapevines grown in microplots from 2012 to 2015. Main effect of P supply (bar charts) on fruitfulness and berry number per cluster (n = 16) and interactive effect of year and P supply (scatter charts) on yield and berry mass (n = 4). A * to the right of a symbol in interactive plots indicates those P treatments that differed from the Control (100% P) in each year based on Tukey s honest significant difference (HSD) at 95% confidence. Data from 2011, before P supply was manipulated, are shown as a reference in interactive plots. Letters above means in bar charts indicate treatment differences based on Tukey s HSD at 95% confidence. Data are means and standard error of the mean for each plot.

Phosphorus in Pinot noir 357 should be based on vine yield (or growth) responses. It is likely that certain yeast strains may have different P requirements than others (similar to variability in YAN needs), but this is presently unknown due to a lack of published data on P requirements in grape musts. Aside from the impact of decreasing P supply on alcoholic fermentation, the effect on other must chemistry parameters must be taken into account. In this study, decreasing P supply in the vineyard led to lower titratable acidity in musts, and increased ph. Lower titratable acidity was also observed in must from Chenin blanc grapevines grown at lower P fertilization rates, although ph was not altered in that trial (Conradie and Saayman 1989b). It is not clear how consistent this effect of P limitation may be in vineyards, since some previous experiments did not examine must acidity (Tulloch and Harris 1970, Skinner et al. 1987, 1988). However, the decreased titratable acidity occurred because P in the must contributes to the pool of titratable acids, either as phosphoric acid or as other metabolites such as phosphorylated sugars and organic acids (Boulton 1980). In peach fruit, P accounted for ~12% of the total titratable protons when titrating to a ph endpoint of 8.2 (Lobit et al. 2002). The changes in titratable acids and ph in response to P supply here were not due to antagonism with K uptake, since must K concentrations were not altered by P supply. The impact of the observed changes in titratable acidity are likely of minor significance, since ph only increased by a tenth of a unit in the no-p vines where yield was already affected. In addition, must ph was not altered enough in any single year to be significant. Prior work showed that the phenolic composition and aroma profiles of Pinot noir berries were not altered in vines where P status was reduced to a similar level as obtained in this study, although wines were not produced in that trial (Schreiner et al. 2014). Berry volatile analysis from this trial showed no consistent impact of P supply on a wide array of aroma compounds (Yuan et al. 2018). For these reasons, we suggest that vine P guidelines for Pinot noir, and likely other cultivars, be based on growth and yield responses to P status, i.e., P limitation is a viticultural production issue and not a winery issue. Both canopy size by veraison (leaf area) and yield were similarly sensitive to P limitation, as both of these parameters were reduced only after three years in vines receiving no P. Leaf area at veraison was reduced in year three by 23% and in year four by 28%, while yield was reduced in year three Figure 4 Effect of year and phosphorus (P) supply on must parameters of Pinot noir grapevines grown in microplots from 2012 to 2015. Main effect (bar charts) of year on must soluble solids (n = 16), or main effect of P supply on must ph and titratable acidity (n = 16), and interactive effect of year and P supply (scatter chart) on must P concentrations (n = 4). Letters near each symbol for must P concentrations designate significant mean groups based on Tukey s honest significant difference (HSD) at 95% confidence and data from 2011, before P was manipulated, are shown as a reference. Letters above means in bar charts indicate year or treatment differences based on Tukey s HSD at 95% confidence. Data are means and standard error of the mean for each plot. Must P data were log-transformed prior to analysis of variance.

358 Schreiner and Osborne by 23%, and in year four, by 35%, compared to the Control vines. These findings support our previous work in ownrooted Pinot noir, where the lowest P supply in that trial in a sand medium never affected growth or yield (Schreiner et al. 2013), and explain why P limitation is generally uncommon in vineyards. Yield responses to P fertilizer (P limitation) have rarely been reported in vineyards, and it is difficult to compare leaf blade or petiole P status from these studies to our findings here. For example, yield of Syrah (Shiraz) vines increased after annual P fertilizer additions for eight years in a sandy loam soil in Australia, but P status of the vines was only measured after 20 years in that trial (Tulloch and Harris 1970). The unfertilized vines had bloom petiole P concentrations of 1.57 g P/kg DW, while P-fertilized vines were 4.55 g P/kg DW after 20 years. Bloom petiole P values in both the 20% P and no-p vines here were below 1.50 g P/kg DW in 2014 and 2015, but yield was only reduced in the no-p vines. Large yield increases of 50 to 70% were found after applications of P in North Coast and Sierra Foothill vineyards in California in vines that were deficient in P (Skinner et al. 1987, 1988). Unfortunately, in these studies, vine P status was measured using acetic acid extracts instead of total P in leaf blades or petioles, making comparisons to the present study difficult. Finally, a small increase in yield (<5%) with P fertilizer additions was observed in Chenin blanc in South Africa by pooling data over 11 years, although individual years were not significant (Conradie and Saayman 1989a, 1989b). Leaf blade P at veraison in that trial was raised from 1.23 g P/kg DW in the non-fertilized vines to 1.33 g P/kg DW in the P-fertilized vines. These values are considerably higher than leaf blade P measured here (0.82 to 0.90 g P/kg DW), when yield was reduced in Pinot noir. These contrasting levels of leaf P where yield was altered indicates that Pinot noir likely has a lower P requirement than Chenin blanc. Indeed, research in California showed that Chardonnay had a lower P requirement than Chenin blanc to obtain maximum yield (Skinner et al. 1988). The notion that Pinot noir may require less P than other cultivars is supported by the observation that Pinot noir consistently has lower petiole P concentrations than other cultivars grown in South Australia (Robinson 2005). We did not observe a clear impact of P limitation on flowering or fruit set, as previously observed (Skinner and Matthews 1989). Indeed, the yield decline reported here in Pinot noir was driven more by reduced berry weight than by a lower number of berries per cluster, and P limitation did not alter flower number per cluster or fruit set in any year. This is consistent with our prior trial with own-rooted Pinot noir, where low P supply also did not affect flowering or fruit set (Schreiner et al. 2013). The present study was extended another year, in part to see if flower number or fruit set would eventually be reduced as a consequence of P limitation. Since they were not, one can conclude that P requirements for flower development and fruit set are lower than P requirements for canopy growth and yield in Pinot noir. Whether or not other grapevine cultivars also have a lower sensitivity to P limitation for flowering parameters versus canopy or berry growth is not known. More systematic research comparing cultivars and possibly even clones among cultivars to examine P requirements for flowering, canopy growth, and yield parameters may be warranted, especially in regions where soil P availability is low. Our finding of lower Mg concentrations in leaves, petioles, and musts in response to low P supply in Pinot noir supports earlier reports that grapevines have difficulty taking up and transporting Mg to aboveground plant parts when P status becomes limiting (Skinner and Matthews 1990). Leaf blade P concentrations measured at veraison resulted in distinct differences among each of the four P supply rates given to vines by 2014 and 2015. In addition, the relative decline in leaf blade P concentrations at veraison as P supply decreased corresponded well with the level of P that vines received. This was not true for petioles at bloom or veraison, or for leaf blades at bloom. P concentrations in all three of these samples showed the largest decline as P supply dropped from the Control (100% P) to the 50% P supply rate, with much smaller declines as P supply was further reduced. In these same samples, the 50% P vines usually did not differ from the 20% P vines, and the 20% P vines never differed from the no-p vines. However, leaf blade P levels at veraison were lower in the no-p vines than in the 20% P vines in 2014 and 2015, when yield was also reduced in no-p vines. Thus, veraison leaf blade P concentrations better diagnose the P status of Pinot noir. That leaf blades at veraison are more diagnostic than leaf blades at bloom or petioles at either time point was not clear when we examined the raw (individual) plot data from this trial previously, using regression to examine vine growth and yield responses to P (Schreiner and Scagel 2017). Results from that analysis showed that leaf blades and petioles did not differ significantly in predicting leaf area at veraison or yield based on P concentrations, although leaf blades had numerically better regression model parameters than petioles (Schreiner and Scagel 2017). By focusing on the treatment means over time here, it is obvious that leaf blades at veraison were superior in diagnosing P status where canopy growth or yield declines occur. Whether or not this result would be the same in irrigated or dryland Pinot noir vineyards requires further testing. Conclusions Reducing P supply in grafted Pinot noir grapevines had the greatest impact on must P concentrations, but the level of P obtained in musts was not low enough to alter yeast fermentation. P limitation did not influence flowering or fruit set in Pinot noir. Tissue guidelines for P status should therefore be based on growth and yield responses, and these were more closely related to leaf blade P concentrations at veraison than to leaf blades at bloom, or to petioles at bloom or veraison. Canopy size based on leaf area at veraison and yield were both reduced in the last two years of this trial, when veraison leaf blade P concentrations were below 1.0 g P/kg DW. Indeed, in the year before yield or leaf area was reduced, veraison leaf blade P in the no-p vines was 1.0 g P/kg DW. This value represents the critical level of leaf blade P at which Pinot noir producers need to increase vine P status. Critical

Phosphorus in Pinot noir 359 levels of P in other tissues are not as reliable as leaf blades at veraison, but suggested levels at which P limitation is likely to occur based on this study are 1.7 g P/kg DW in bloomtime leaf blades, 1.2 g P/kg DW in bloomtime petioles, and <0.5 g P/kg DW in veraison petioles. Minimum yeast P requirements for Pinot noir musts are below 32 mg P/L. Literature Cited Archer TE and Caster JGB. 1956. Phosphate changes in fermenting must in relation to yeast growth and ethanol production. Am J Enol Vitic 7:62-68. Bisson LF. 1999. Stuck and sluggish fermentations. Am J Enol Vitic 50:107-119. Boulton R. 1980. The relationships between total acidity, titratable acidity and ph in grape tissue. Vitis 19:113-120. Conradie WJ and Saayman D. 1989a. Effects of long-term nitrogen, phosphorus, and potassium fertilization on Chenin blanc vines. I. Nutrient demand and vine performance. Am J Enol Vitic 40:85-90. Conradie WJ and Saayman D. 1989b. Effects of long-term nitrogen, phosphorus, and potassium fertilization on Chenin blanc vines. II. Leaf analyses and grape composition. Am J Enol Vitic 40:91-98. Cook JA. 1966. Grape nutrition. In Nutrition of Fruit Crops. Childers NF (ed.), pp. 777-812. Somerset Press, Somerville, NJ. Dukes BC and Butzke CE. 1998. Rapid determination of primary amino acids in grape juice using an o-phthaldialdehyde/n-acetyll-cysteine spectrophotometric assay. Am J Enol Vitic 49:125-134. Fujitani T. 1965. Biochemical studies on the mineral components in sake yeast part 2. The requirements of phosphorus, potassium and magnesium by the yeast. Agr Biol Chem 29:477-485. Hoagland DR and Arnon DI. 1950. The Water-Culture Method for Growing Plants without Soil. California AES Circular, Berkeley, CA. Jones JB and Case VW. 1990. Sampling, handling, and analyzing plant tissue samples. In Soil Testing and Plant Analysis, 3rd ed. Westerman RL (ed.), pp. 389-427. Soil Science Society of America, Madison, WI. Jones RP and Gadd GM. 1990. Ionic nutrition of yeast physiological mechanisms involved and implications for biotechnology. Enzyme Microb Tech 12:402-418. Klein I, Strime M, Fanberstein L and Mani Y. 2000. Irrigation and fertigation effects on phosphorus and potassium nutrition of wine grapes. Vitis 39:55-62. Lobit P, Soing P, Génard M and Habib R. 2002. Theoretical analysis of relationships between composition, ph, and titratable acidity of peach fruit. J Plant Nutr 25:2775-2792. Robinson JB. 2005. Critical plant tissue values and application of nutritional standards for practical use in vineyards. In Proceedings of the Soil Environment and Vine Mineral Nutrition Symposium. Christensen LP and Smart DR (eds.), pp. 61-68. ASEV, Davis, CA. Sampaio TL, Kennedy JA and Vasconcelos MC. 2007. Use of microscale fermentations in grape and wine research. Am J Enol Vitic 58:534-539. Schreiner RP and Scagel CF. 2016. Arbuscule frequency in grapevine roots is more responsive to reduction in photosynthetic capacity than to increased levels of shoot phosphorus. J Amer Soc Hort Sci 141:151-161. Schreiner RP and Scagel CF. 2017. Leaf blade versus petiole nutrient tests as predictors of nitrogen, phosphorus, and potassium status of Pinot noir grapevines. HortScience 52:174-184. Schreiner RP, Scagel CF and Baham J. 2006. Nutrient uptake and distribution in a mature Pinot noir vineyard. HortScience 41:336-345. Schreiner RP, Zasada IA and Pinkerton JN. 2012. Consequences of Mesocriconema xenoplax parasitism on Pinot noir grapevines grafted on rootstocks of varying susceptibility. Am J Enol Vitic 63:251-261. Schreiner RP, Lee J and Skinkis PA. 2013. N, P, and K supply to Pinot noir grapevines: Impact on vine nutrient status, growth, physiology, and yield. Am J Enol Vitic 64:26-38. Schreiner RP, Scagel CF and Lee J. 2014. N, P, and K supply to Pinot noir grapevines: Impact on berry phenolics and free amino acids. Am J Enol Vitic 65:43-49. Schreiner RP, Osborne J and Skinkis PA. 2018. Nitrogen requirements of Pinot noir based on growth parameters, must composition, and fermentation behavior. Am J Enol Vitic 69:45-58. Skinner PW and Matthews MA. 1989. Reproductive development in grape (Vitis vinifera L.) under phosphorus-limited conditions. Sci Hortic 38:49-60. Skinner PW and Matthews MA. 1990. A novel interaction of magnesium translocation with the supply of phosphorus to roots of grapevine (Vitis vinfera L.). Plant Cell Environ 13:821-826. Skinner PW, Matthews MA and Carlson RM. 1987. Phosphorus requirements of wine grapes: extractable phosphate of leaves indicates phosphorus status. J Amer Soc Hort Sci 112:449-454. Skinner PW, Cook JA and Matthews MA. 1988. Responses of grapevine cvs Chenin blanc and Chardonnay to phosphorus fertilizer applications under phosphorus-limited soil conditions. Vitis 27:95-109. Toh-e A, Ueda Y, Kakimoto SI and Oshima Y. 1973. Isolation and characterization of acid phosphatase mutants in Saccharomyces cerevisiae. J Bacteriol 113:727-738. Tulloch HW and Harris WB. 1970. Fertilizer responses with nonirrigated Shiraz grapevines, 1944-1966. Aust J Agric Res 21:243-252. Yuan F, Schreiner RP and Qian MC. 2018. Soil nitrogen, phosphorus, and potassium alter β-damascenone and other volatiles in Pinot noir berries. Am J Enol Vitic 69:157-166.