Humidity tents

Humidity tents consisted of 90 × 57 × 22 cm (35 × 22 × 9 in) molded plastic utility tubs, into which we had drilled drainage holes in the bottom. We fit them with 1.8 m x 1 cm (6 ft x 0.5 in) PVC pipes bent into arches, one at each end of the tubs plus a third in the middle, to hold up white 70% shade plastic sheeting. We folded the plastic tightly at the corners and held it in place on the sides and ends with large paper clamps to form a moderately tight seal. These humidity tents can accommodate 180 cuttings each, spaced 4 by 6 cm (1.5 by 2 in) apart. They can accommodate cuttings up to 75 cm (30 in) tall at the centers of the tents, but only about 45 cm (18 in) tall at the perimeters. Materials cost about $30 per humidity tent.

We placed the tubs in a greenhouse with temperatures set at 21 C (70 F) day and 18 C (64 F) night under halogen lamps set for 16 hr days. We filled the tubs to a depth of 20 cm (8 in) with a 1:4 mixture of peat and perlite, then soaked it and allowed it to drain. For Experiment 1, we kept the tents sealed at the sides, to maintain humidity near 100% for the duration of the trial. For Experiment 2, we kept the tents sealed until greenhouse temperatures started to rise in the spring, at which time we opened them partially and added humidifiers to maintain high humidity. If humidity fell below 60%, we watered the rooting medium until water dripped out the drainage holes. High-low temperature and humidity remote sensors (Control Company, Traceable® Products, 12554 Galveston Road Suite B230 Webster, TX 77598) placed within the tents showed that temperatures fluctuated between 13 C and 44 C.

Plants

“Crown suckers” are long straight first year suckers that emerge directly from the ground, usually in response to coppicing. Whereas layering is implemented on crown suckers in early summer when they are still green and relatively fragile, for hardwood cuttings the suckers are left until they go dormant in the fall. By that time, the suckers for these experiments had developed a brown lignified bark, and range in length from 40 to 190 cm (16 to 75 in) and in basal diameter from 3 to 17 mm (0.1 to 0.7 in). For these experiments, suckers were harvested soon after leaf drop in late October, at the same time as rooted layers are typically dug. They were cut as close to the ground as possible, bundled by genotype, and stored in a 2 C (36 F) cooler maintained at high humidity until preparation for rooting.

We conducted these trials early in our breeding program, at a time when we still had no clonally produced mother plants, nor any named varieties. The suckers used for these trials were regrowth from mother plants that had been mound layered a year previously to produce plants for evaluation in replicated trials. With only a single mother plant per genotype, we were limited by the number of stems per genotype, which was highly variable. Moreover, stem size was also highly variable within each genotype. To maximize statistical significance, we used as many stems as we could obain, and included genotype in the statistical models to accommodate the resulting unbalanced experimental design.

Stem Preparation

During preparation, we handled suckers quickly to prevent dehydration. We trimmed about 1 cm (0.5 in) off the base of each sucker before cutting it into the segment lengths desired for each experiment. Then we dipped about 2 cm (1 in) of the base of each segment into a solution of indole-3-butyric acid (IBA) dissolved in a 1:1 mix of ethanol with deionized water, for 10 to 15 seconds. The concentration of IBA varied according to the treatment. Finally, we inserted the segments into the rooting medium, only as deeply as needed for the cutting to stand vertically, about 3 to 5 cm (1 to 2 in). As soon as each humidity tent was full, we watered it heavily and sealed it closed with plastic. After that, we only opened humidity tents briefly to check on humidity and water as needed, and for evaluations.

Evaluation and Care after Rooting

We evaluated root formation at three to four week intervals, starting about two months after preparation, with up to six evaluations. During evaluations we used a spritz bottle to keep leaf surfaces wet to reduce moisture stress because the shoots that had developed within the tents were adapted to the high humidity conditions within the tents and wilted quickly when exposed to drier ambient conditions.

To evaluate, we gently pulled each cutting from the growing medium. We recorded which cuttings were dead and discarded them. We returned cuttings that had not rooted to the same position in the humidity tent. We gave cuttings that had rooted a subjective rating of root quality based on length, thickness and number of roots, and of leaf quality based on number, size and color of leaves. We did not take detailed measurements of root number and root length because transplant survival data was important and thus we did not want to risk moisture stress to the tender roots. We then transplanted rooted cuttings into 8 cm (3 in) square pots with a 3:1 mixture of a commercial bark-based potting mix plus perlite, plus 5 g^.L⁻¹ (5,000 ppm) Osmocote™ 15-9-12 + micronutrients.

We placed the newly transplanted rooted cuttings inside another humidity tent to harden off under the same high initial humidity conditions to which they were adapted. After two or three days the rooted and potted cuttings were gradually acclimated to ambient greenhouse conditions by partially opening the plastic for an increasing number of hours each day. As conditions allowed, we transitioned them outdoors, then transplanted them to the field in mid- to late September.

Experiment 1: IBA concentration and segment position

We collected crown suckers from 22 genotypes in fall 2009, from 6 to 24 suckers per genotype, and stored them until March 16, 2010, when we prepared them for rooting. We also included a few first-year canopy stems for comparison, as well as a few second-year suckers, which could be identified by their slightly more lignified bark and a few catkins.

We divided 326 stems into 968 segments. We prepared one genotype at a time. First, we sorted each bundle of stems into groups of six stems of approximately the same length, which we treated as an experimental block. We then cut these six stems into 20 cm (8 in) long segments, which we kept grouped by segment position: basal, medial 1, medial 2, medial 3, etc., terminal. Most stems were long enough to be divided into at least two or three segments; a few were long enough to divide into as many as five or six, or even seven. A few were too short to be divided at all and were labeled “undivided”.

Within each genotype by segment-type grouping, we randomly assigned segments to one of six IBA treatments: 0, 1, 2, 4, 8 and 16 g L⁻¹ (0, 1,000, 2,000, 4,000, 8,000, and 16,000 ppm). For the controls, we dipped the cuttings in plain 1:1 ethanol/DI water. After segment bases were dipped in the IBA solution, we inserted them in the rooting medium in a rectangular array, with IBA concentration on one axis and segment-position on the other. We controlled for possible microclimate variation within the humidity tents by randomizing the orientation of each grouping relative to that of other groupings. We evaluated root formation 56, 79, 101 and 122 days after the stems were placed in the tents, that is, approximately every three weeks from May 11 through July 16. We recorded survival of the rooted and potted cuttings on September 30, six and a half months after the cuttings were first prepared for rooting.

Experiment 2: Stem size and stem segment size

We collected 149 crown suckers of five different genotypes, from 12 to 32 stems per genotype, in late fall 2013, and stored them in a cooler until Jan. 22, 2014. These stems ranged from 43 to 155 cm (17 to 61 in) long and from 2.5 to 12 mm (0.1 to 0.5 in) wide at the base. The first step to prepare them was to identify stems of similar size for blocking. Working with one genotype at a time, we selected the five longest stems and randomly assigned them to be cut either into six, five, four, three or two segments of even length. In this way, we produced groups of segments ranging in length from 15 to 75 cm (6 to 30 in), all from stems that were initially of similar size. Then we repeated this process with the next five longest stems. If there were not five stems long enough to be cut into six segments of at least 15 cm (6 in), we selected just four stems, and cut them into five, four, three and two segments each, and so forth. When the only remaining stems were short enough to fit into the humidity tents without division (75 cm (30 in) or shorter) we cut them into three or two pieces of even length, or left them undivided. We prepared 429 cuttings in this way.

Before cutting the stems, we measured the length, basal diameter and weight of each. After cutting them, we measured the length, basal diameter and weight of each segment, and counted numbers of viable buds on each segment. Finally, we dipped their bases in 2 g^.L⁻¹ (2,000 ppm) IBA and inserted them into the rooting medium in the humidity tents, in order of segment position, in order to keep track for record keeping.

We evaluated root formation 72, 98, 121, and 149 days after the start of the experiment, with June 19 being the last evaluation date. We recorded survival of the rooted and potted cuttings on September 15, 2014, eight months after the cuttings were first prepared for rooting and just before field transplanting.

Statistical Analysis

For Expt. 1, we used JMP software (v. 12, Copyright 2015, SAS Institute, Inc.) for ANOVA and means comparisons. For Expt. 2, we used XLISP-STAT software (version 3.52.17, Copyright 1989-1999, by Luke Tierney) for regression analysis. We used binomial regression for rooting and survival data (rooted = 1, unrooted = 0), and linear regression for root and leaf quality ratings. We considered single cuttings as the experimental units and controlled for other factors, such as genotype, by including them as covariates in the statistical models.

Experiment 1: IBA concentration and segment position

Root formation was highly variable by genotype (data not shown). Some genotypes calloused heavily, but callous formation was not related to rooting success. Stems from bushes that rooted well in mound layering rooted at higher rates than from bushes that did not (p = 0.0001). Stems from bushes that had been coppiced for mounding rooted at a higher rate than those that had not been coppiced (p < 0.01). Canopy stems did not root at all; neither did second year suckers. Basal segments and undivided stems rooted at higher rates than medial or terminal segments at p < 0.0001 (Table 1).

Table 1 Percent root formation for various IBA concentrations by position of segments cut from hardwood stems of hybrid hazelnuts149 days after the start of the experiment. Means within a row that share an upper case letter do not differ at Tukey HSD α = 0.05. Means within a column that share a lower case letter do not differ at Tukey HSD α = 0.05

View Fullscreen

Although a few cuttings rooted with no IBA, root formation was significantly higher with 1, 2, and 4 g^.L⁻¹ (1,000, 2,000, and 4,000 ppm) IBA than with either no IBA or 16 g^.L⁻¹ (16,000 ppm) IBA (p < 0.0001) (Table 1). The number of days to root was significantly reduced with higher concentrations of IBA (p = 0.0003). That is, rooting occurred faster with higher concentrations of IBA. However, higher concentrations of IBA also correlated with higher cutting mortality at the first evaluation (p = 0.0001).

Root quality was not affected by IBA concentration, but increased significantly with increasing segment length (p < 0.0001). Basal segments and undivided stems produced more abundant and longer roots than either terminal or medial segments (p < 0.0013 and p = 0.0032 respectively). Survival of the first set of potted cuttings, which were potted in mid-May, was 91% when assessed in late August (data not shown). Survival of subsequent sets was lower, down to 75% for those potted in mid-July, but still much higher than the survival rates we observed with softwood cuttings. We did not trace survival to previous treatment.

Genetic variability in ability to root, as supported by our observation that the genotypes that rooted well in layering also tended to root well from cuttings, has been well documented in hazelnuts (Kreiser et al. 2016). The inability of canopy stems and two-year-old suckers to root is probably related to the loss of juvenility and loss of rooting competence associated with maturation, as has been noted in many species (Bellini et al. 2014).

The concentrations of IBA that we found to be the best were in the same range as identified by others for “quick dip” application for other woody plants (Hinesley and Blazich 1981). However, Rodrigues et al. (1988) found that when hardwood hazelnut cuttings were soaked in a 70% ethanol IBA solution for 24 hrs, 0.05 g^.L⁻¹ (50 ppm) was sufficient. Whereas many authors found a linear increase in rooting with increasing concentrations, others (Blythe and Sibley 2009) found that shoot growth was reduced with high auxin levels, with deleterious effects (Chadwick and Burg 1967). This is consistent with our finding that root formation was inhibited at 16 g^.L⁻¹ (16,000 ppm).

We had previously found between 1 and 2 g^.L⁻¹ (1,000 and 2,000 ppm) to be the best IBA concentration for softwood hybrid hazelnut cuttings (unpublished data). We hypothesized that higher concentrations might be necessary for larger basal segments, which are more lignified, to compensate for poor penetration of the IBA solution into the stem cuticle. The data support but do not confirm this hypothesis: the highest rooting was attained with concentrations of 1 to 2 g^.L⁻¹ (1,000 and 2,000 ppm) for terminal segments, 2 g L⁻¹ (2,000 ppm) for medial segments, and 2 to 8 g^.L⁻¹ (2,000 to 8,000 ppm) for basal segments. However, because results for 2 g^.L⁻¹ (2,000 ppm) did not differ significantly from other concentrations for any segment type, we chose 2 g^.L⁻¹ (2,000 ppm) as the IBA concentration for all future hardwood stem cutting trials.

Experiment 2: Stem size and stem segment size

As in Experiment 1, root formation was highly variable by genotype. Up to 42% of segments from the best three genotypes formed roots, but fewer than 4% of the two worst genotypes rooted. We excluded these two genotypes from further statistical analysis.

As in Experiment 1, the highest rooting percentage was attained with undivided stems as 69% of undivided stems rooted, of which 91% survived the transition to pots, meaning that 63% of all undivided stems produced viable new plants. These were stems that ranged from 46 to 75 cm (18 to 30 cm) long, had basal diameters from 2.5 to 6.1 mm (0.1 to 0.5 in), and weighed from 2 to 9 g (0.07 to 0.32 oz). For these stems, which were small relative to the rest of the stems, there were no significant differences in root formation or survival of potted cuttings due to any stem size parameter.

Rooting percent was negatively correlated with the diameter of the source stem (Fig. 1). In other words, thick stems rooted less frequently than thin ones. This trend was evident for all segment types (basal, medial, and terminal), but not for undivided stems, but was only statistically significant for medial segments (p < 0.003) and for the aggregate of all segment types (p = 0.0002). The ratio diameter/length was also negatively correlated with rooting success (p < 0.0001), indicating that rooting success is higher with long thin stems than with short thick ones. The diameters of stems that formed roots ranged from 2.5 to 9.5 mm (0.1 to 0.4 in), with the majority of rooted segments coming from stems between 6 and 8.5 mm (0.2 to 0.3 in) wide, which were the most common size of stems harvested. Stem length, weight and bud number had no effect on rooting.

View Figure

• Download as Powerpoint
• Download Figure

Unlike in Experiment 1, in Experiment 2, there were no significant differences in root formation between basal, medial and terminal segments. Of all segment parameters, segment length was the most significant predictor of rooting. Rooting percent increased linearly with increasing segment length (Fig. 2). This trend was evident for all segment types except for medial, but only statistically significant for basal segments (p < 0.04) and for the aggregate of all segment types (p = 0.0003). (Segment length was equal to stem length in the case of undivided stems.) Conversely, rooting decreased linearly with increasing segment basal caliper (p < 0.05). Segment weight had no effect on rooting. Thicker and heavier segments survived in the humidity tents longer without rooting than thinner and lighter segments (p < 0.0001), whereas segment length had no effect on survival of unrooted cuttings.

View Figure

• Download as Powerpoint
• Download Figure

The inverse relationship between rooting ability and stem diameter that we observed in hazelnut stem cuttings has been noted in many other genera, including Syringa and Prunus (Howard, 1994) and Acacia (Schwarz, et al. 1999). These authors speculated that as stems grow thicker, either increased sclerenchymatous fiber or increased suberization of the cortex present a physical barrier to emergence of adventitious roots. This is more of an issue with difficult-to-root species, such as hazelnuts, than with easy-to-root species. Several authors (Moe 1988, Howard 1994, Murray et al. 1994) have noted that various methods of pre-treatment of stock-plants before cutting collection, such as heavy pruning, etiolation and blanching, seem to enhance rooting by keeping stems juvenile and by inhibiting growth in stem diameter, whereas pre-collection treatments that allow stems to grow normally do not improve rooting. This explains our observation that sometimes stems that failed to root in mound layering, stems with blanched basal sections, can later be rooted as stem cuttings. We subsequently found that blanching stems while they are developing in the stock plant beds will enhance their later rooting from cuttings (unpublished data).

The positive correlation between rooting and segment length that we observed is likely to be simply a function of increased carbohydrate stores in longer cuttings. The carbohydrate status of cuttings has been shown to affect the rooting ability of many species, including Populus (Friend et al. 1994, Kaczmarek et al. 2014), and Persea, Rubus and Malus (Hartmann and Kester 2011). This makes sense because carbohydrates are essential for providing both the energy and the carbon skeletons needed for production of root tissue.

However, as Veierskov (1988) noted, carbohydrate stores only improve rooting if other prerequisites for rooting are also present. Thus, although carbohydrate stores also increase with increasing diameter, this is negated by increasing lignification of thicker stems. The trade-off between the inhibition of root emergence in stems with a thick cortex and the contribution of carbohydrate stores to root development explains why segment weight was not as predictive of rooting as either segment diameter or length, even though it is a mathematical product of both: the positive of one neutralizes the negative of the other. Conversely, that the best rooting was attained with undivided stems probably reflects that these stems represented the best trade-off between these opposing trends: stems that were short enough to fit in the humidity tents without cutting also tended to be thinner than the others, but because they were not subdivided, they had access to their full length of carbohydrate stores. It is also possible that mobile carbohydrates were not evenly distributed in stems, as a function of differences in rates of basipetal translocation, in which case when a stem was divided some segments may have ended up with disproportionately more or less carbohydrate. That might explain why usually only one section from a stem would form roots. Endogenous IAA from terminal buds may also have augmented rooting of the undivided stems (Hartmann and Kester 2011).

Carbohydrate stores also explain why thicker and heavier segments could stay alive without rooting for much longer than thinner and lighter weight segments, which would have depleted their carbohydrate stores more quickly through respiration. They also explain why survival after rooting was improved with all parameters of segment size, including diameter and weight: after the root primordia succeed in penetrating the stem cortex, what matters is having enough carbohydrates to grow those roots into functioning organs of water and nutrient uptake.

In Expt. 1, in which we cut stems into many relatively short uniform lengths, we found that basal segments formed roots at higher rates than medial and terminal segments. This was probably a function of the larger carbohydrate stores in basal segments, due to the natural taper of stems. The combined results of Experiments 1 and 2 suggest that if stems are to be divided, they should be divided in such a way as to ensure even distribution of stored carbohydrates between segments. In other words, stems should be divided so that terminal segments are longer than basal segments to ensure that both get adequate carbohydrate stores. We have since tested this theory in another trial and found it to have merit.

Root quality increased with all parameters of segment size at p < 0.05 in terminal segments, but root quality was not predicted by either stem size or segment size in other types of segments or in undivided stems. By contrast, leaf quality improved with increasing stem caliper (p < 0.001) and with increasing stem weight (p <0.05), but not with stem length. For subdivided stems of all segment types, leaf quality improved with increasing segment caliper at p < 0.001 and with increasing segment length and weight at p <0.0001. Moreover, leaf quality was highly correlated with root quality (p = 0.0001).

The best predictor of survival of rooted and potted cuttings, measured in mid- to late September, was leaf quality, rated subjectively at the time of potting (p = 0.003). Whereas overall survival was 63%, survival of rooted cuttings that had healthy leaves and actively growing shoots at time of potting was 85%. By contrast, root quality, also a subjective rating, predicted survival at only p = 0.07. Survival of rooted and potted cuttings was also positively correlated with most parameters of segment size (p < 0.03), as expected due to the correlation between those parameters and leaf quality. Although we did not trace field survival to previous treatment, overall field survival averaged 89%, though it varied strongly by genotype; field survival for the best genotypes was 100%.

We were surprised that leaf quality at the time of potting was a better predictor of future survival than root quality. Davis (1988) and Dick and Dewar (1992) hypothesized that shoot and root development compete for limited carbohydrate stores. Such competition is plausible with these cuttings, which broke dormancy within two to three weeks of preparation and quickly developed leafy shoots that were up to 10 cm long, whereas we did not observe much root development until after three to four months. However, if the new leaves are photosynthetically functional, the carbohydrates they produce may replace the carbohydrates used in their growth, and increase the amount available for root growth. The contribution of current photosynthesis to rooting and survival after rooting has been debated, but neither proven nor disproven (LeBude et al. 2005).

To reduce competition between shoots and roots, Loach (1988) recommends inhibiting shoot growth during rooting by keeping shoot temperatures cooler than root temperatures. In an unpublished experiment, we tested this hypothesis by trying to root cuttings on a heating mat inside a dark cooler, using a layer of rock wool mulch on the surface of the medium to keep the medium moist and warm, without also warming the shoots. The only growth that occurred, besides callous, was of highly etiolated vegetative shoots that emerged from buds below the warmed rooting medium. The lack of root development suggested that leaf development or photosynthesis may indeed have been needed for root developing in these hazelnut cuttings, as found for Populus robusta by Wareing and Smith (1963, cited by Roberts and Fuchigami 1973).

A key question is whether leaves are functional in humidity chambers. Some authors report evidence that they are not, due to moisture stress and due to inundation of stomata by water droplets, which impedes CO₂ diffusion, especially in mist systems (Davis 1988). Neither of these were likely to have been problems in our system because we did not use mist, and because the leaves did not show signs of moisture stress. It is likely that the leaves were highly adapted to the humidity conditions of the tents, because they had emerged and developed under those conditions. We observed that leaf quality and root quality were highly correlated. In fact, we could often identify a cutting with large roots, even before pulling it from the rooting medium, by the quality of its leaves. This suggests not only that the leaves were photosynthetically functional, but that they supplied valuable carbohydrates for root growth after root initiation. This is supported by a previous trial (unpublished) in which we found that supplemental light improved survival of rooted cuttings.

A purported advantage of propagation from stem cuttings over mound layering is that theoretically more than one new plant can be produced from each stem. In reality, subdividing stems may not be productive. Of the 80 stems that were subdivided into multiple shorter segments, only 25% produced two or more rooted segments, whereas 44% produced just one rooted segment, and the remaining 31% produced none, for a total of only 92 rooted. It may thus appear that subdividing stems produced 12 more rooted segments than would have been obtained without subdividing them. But that is not a valid assumption: given that undivided stems rooted at higher rates than divided ones, some of the divided stems that failed to produce any rooted segments may have rooted if they had been kept whole. So we cannot conclude that the practice of dividing stems produces more rooted plants. We can only recommend that stems be divided only as necessary to fit them into the rooting chamber. The only drawback to doing so is that the tall rooted cuttings that result tend to be very top-heavy and thus require staking in their pots until they grow enough roots to anchor them well.

We had hoped that our results would point to an optimum size for sucker diameter and segment length, but most responses were linear for the range of sucker sizes tested: rooting was best the thinner the crown sucker and the larger the segment into which suckers were cut. Our recommendations are as follows. 1) Choose healthy suckers that are as thin as possible, but no thinner than 3 mm in diameter, as we did not test any smaller than 3 mm, 2) Leave them undivided if possible. 3) However, if they must be cut to fit into the humidity tents, keep them as long as possible, and no shorter than 25 cm, and try to allocate biomass evenly between the terminal and basal segments. This has been our standard procedure since we completed this trial; subsequent rooting success has supported this approach.