Preliminary Species and Media Selection for the Veggie Space Hardware

Pillows were constructed from 7.6 cm x 12.7 cm bags (3″ x 5″ re-closable static shielding bags, Uline, Pleasant Prairie, WI, external dimensions of 10 cm x 15 cm). On one side, a 5.3 cm x 10.5 cm window was cut, and a 7.5 cm x 11.5 cm piece of Nitex nylon mesh (Sefar Nytal PA-25-63, Sefar, Heiden, Switzerland) was heat welded to this surface to allow capillary wicking. Each pillow was filled with 100 mL of media and fertilizer (discussed below) and sealed using the re-closable zipper. In the non-wicking surface of the pillow, 1.5 cm slits were placed, and two 1.5 cm x 5.5 cm Nitex wicks were passed into each slit so that half of each wick protruded and the other halves were spread along the surface of the media in the pillow. Wicks were used to enclose seeds and promote early germination events by maintaining moisture around the seed.

The first cultivar and media test examined 13 cultivars in eight species: lettuce (Lactuca sativa L.) cultivars ‘Outredgeous’ (Johnny’s Select Seed, Winslow, ME), ‘Flandria’ (Rijk Zwaan USA, Salinas, CA), ‘Sierra’ (Veseys, York, PEI, Canada), and ‘Oak leaf’ (Terroir Seeds, Chino Valley, AZ), mizuna, (Brassica rapa cv. Nipposinica) (Seeds of Change, Santa Fe, NM), ‘Tender leaf’ vegetable amaranth (Amaranthus viridis cv. Tender leaf) (Evergreen YH Enterprises, Anaheim, CA), ‘Bright lights’ Swiss chard (Beta vulgaris var. cicla) (Terroir Seeds, Chino Valley, AZ), ‘Tokyo Bekana’ Chinese cabbage (Brassica rapa, Chinensis group) (Evergreen YH Enterprises, Anaheim, CA), ‘Sugar Pod II’ snow pea (Pisum sativum cv. Sugar Pod II) (Evergreen YH Enterprises, Anaheim, CA), ‘Spicy Globe’ basil (Ocimum basilicum minimum ‘Spicy Globe’) (Ferry-Morse Seed Co., Fulton, KY), ‘Genovese’ basil (Ocimum basilicum ‘Genovese’) (Terroir Seeds, Chino Valley, AZ), common chives (Allium schoenoprasum) (Veseys, York, PEI, Canada), and Greek oregano (Origanum heracleoticum) (Terroir Seeds, Chino Valley, AZ).

All of the leafy greens selected can be grown using the cut-and-come-again strategy, essentially repeated harvests from the same plant, a harvest scenario that will potentially be employed in Veggie. Lettuce cultivars ‘Flandria’ and ‘Outredgeous’ have been previously studied for growth in off-Earth scenarios (Richards et al., 2006; Stutte et al., 2009; Stutte et al., 2011a, b). ‘Sugar Pod II’ snow pea was selected to test a small reproductive species in the pillows. The herbs were chosen to provide different growth habits and with the intent that they could easily be used to add interest and variety to a diet based on pre-packaged thermally stabilized foods.

These 13 cultivars were tested in both Fafard #2, a peat-based commercial potting mix also containing perlite and vermiculite, (Conrad Fafard Inc., Agawam, MA) and in arcillite (sifted 1-2mm, Turface Proleague, Profile LLC, Buffalo Grove IL) using the rooting pillows. The 1-2 mm arcillite medium has been well studied for use in microgravity (Jones et al., 2002; Norikane et al., 2005). Additionally, two cultivars, ‘Outredgeous’ and ‘Sierra’ lettuce were tested with five media types: Fafard # 2, arcillite (1-2mm), a 1:1 blend of the two media, a 7:3 blend (70% Fafard #2: 30% arcillite) and a 1:1 blend of coarse perlite and fine vermiculite (PVP Industries, North Bloomfield, OH). Media were autoclaved and mixed with Nutricote (18-6-8, type 180, Florikan, Sarasota, FL) slow release fertilizer at a rate of 7.5 g/L dry media. Pillows were weighed and then soaked in de-ionized (DI) H₂O for five minutes prior to planting.

Fresh media was soaked in DI H₂O for 70 hours to check pH levels. Starting pH of the media with fertilizer ranged from 5.76 for the Fafard #2 to 6.56 for the arcillite with the blends having intermediate values. The perlite: vermiculite blend was higher at 6.64. Two slit openings were planted in each pillow except for chives, where ten openings were planted per pillow. Two seeds were planted per opening and thinned to one plant per opening upon germination and emergence, except in the case of snow pea where one seed per opening was planted due to large seed sizes. Pillows were placed Nitex-side-down in Veggie reservoir analogs, which consisted of polypropylene sterilizing pans (14.2 L, Thermo Fisher Scientific, Rochester NY) with a sheet of soft foam (43.5 cm x 35.5 cm x 5 cm, Uline, Pleasant Prairie, WI), topped with a 420 mm x 340 mm x 2 mm sheet of PVC and wrapped with a 680 mm x 520 mm piece of Nomex wicking cloth (Aramid fabric, Insulsafe Textiles). PVC was placed over the foam to maintain uniform moisture and eliminate the impact of air bubbles under the cloth. Thirteen pillows were placed in each tub with cultivars randomized in three tubs (two with Fafard #2, one with arcillite). Thirteen pillows of ‘Outredgeous’ or ‘Sierra’ lettuce were placed in each of two other tubs with five media types randomized.

Initially, 2 L of DI H₂O were placed in each tub. Water use was monitored and the level was adjusted daily with DI H₂O to maintain adequate water in the tubs. Tubs were covered with plastic wrap to increase internal humidity until ¾ of the seeds had germinated. Plants were grown under 16-h photoperiods with initially 150 μmol·m^-2·s^-1 triphosphor fluorescent light (Sylvania FP541/841/HO). Light levels were increased to 300 μmol·m^-2·s^-1 at 9 days after planting (DAP). Seeds that failed to germinate were replanted within 5 DAP. Plants were grown at 24°C for the first week and then the temperature was ramped to 28°C day / 24°C night by 9 DAP. Relative humidity was adjusted to 70% day / 75% night and CO₂ was set to 1200 μmol mol^-1 so that environmental conditions were relevant to those measured in the Veggie hardware with bellows closed and lights cycling in a high CO₂ environment like the ISS. All plants were harvested at 36 DAP and root-filled pillows were weighed and characterized at 37 DAP.

Upon harvest, roughly one third of the plants were reserved for microbial load analysis. Plants were photographed and biometric analysis included measurements of chlorophyll, leaf area, fresh and dry mass. Chlorophyll measurements were obtained by averaging the SPAD values of three non-senescent leaves per plant (SPAD-502, Konica Minolta, Osaka, Japan), and average values per pillow were calculated from the two plants. Leaf area was only measured on plants not used for microbial analysis. Shoot tissues were bagged and dried in an oven at 70°C for greater than 72 h. Oven dry tissue was allowed to equilibrate to room temperature and RH for an hour prior to weighing.

Qualitative measurements of root status were obtained by observing pillow fill with roots through the translucent Nitex surface using the following criteria: Pillows filled to excess with roots bulging and solid mats of roots were considered root bound. Pillows ¾ full to full but without bulging were considered full. Pillows ½ to ¾ full were considered half full. Pillows less than ½ full were considered low. Pillows were also weighed and dried, and dry mass was obtained, however pillow-to-pillow media variability and the relative mass of roots to media made it difficult to estimate root dry mass. Arcillite is significantly heavier than dried root tissue, while peat, perlite and vermiculite are similar in mass to dried root tissue.

Plants used for microbial analysis were from one set of all cultivars grown in Fafard #2 and one set each of ‘Outredgeous’ and ‘Sierra’ lettuce grown in three types of rooting media: Fafard #2, 1:1 blend of Perlite and Vermiculite, and arcillite (1-2 mm), described above. The mixes of Fafard #2 and arcillite were not analyzed for microbial load. The two plants grown in one pillow were pooled samples. Shoot tissue was cut from the plants and weighed using aseptic technique and placed into sterile blender bags (BagLight, Interscience Labs, Weymouth, MA). Sterile DI H₂O was added to each sample in a 1:10 weight/volume ratio. Bags containing sample and diluent were placed in a bag mixer (Interscience Labs, Weymouth, MA) and blended for 2 minutes to remove microbes from plant surfaces. The sample extracts were serially diluted in sterile DI H₂O and plated in duplicate onto Difco R2A agar (BD, Franklin lakes, NJ). Plates were incubated at 28°C for 48 hours before enumeration of colonies to determine colony-forming units (CFU) per gram of tissue. Microbial sampling and analysis methods were modified from Pouch Downs and Ito (2001).

Five cultivars from four plant families were tested in the second trial. Cultivars were ‘Cherry Bomb II’ radish (Raphanus sativus L. cv. Cherry Bomb II, family Brassicaceae) (W. Atlee Burpee & Co., Warminster, PA), ‘Sugar Pod II’ snow pea (Pisum sativum cv. Sugar Pod II, family Fabaceae) (Evergreen YH Enterprises, Anaheim, CA), ‘Tokyo Bekana’ Chinese cabbage (Brassica rapa, Chinensis group, family Brassicaceae) (Evergreen YH Enterprises, Anaheim, CA), ‘Bright Lights’ Swiss chard (Beta vulgaris var. cicla, family Chenopodiaceae) (Terroir Seeds, Chino Valley, AZ) and ‘Outredgeous’ lettuce (Lactuca sativa cv. Outredgeous, family Asteraceae) (Johnny‘s Select Seed, Winslow, ME).

Four previously used media were tested: Fafard #2, arcillite (1-2 mm), and 7:3 and 1:1 mixtures of the two of these. Pillows and planting were as described for Trial 1. Sixty pillows were grown in five tubs, therefore there were twelve pillows of each cultivar with three pillows of each media type per cultivar. The sixty pillows were completely randomized within the five tubs, giving twelve pillows per tub arranged in alternating orientation within the tubs to provide maximum spacing for each plant and reduce shading.

Planting and experimental design followed the procedures described above, except that temperature was ramped up to 28°C by 7 DAP and light was maintained at 200 μmol·m^-2·s^-1 for the duration of the experiment. Plants were harvested at 28 DAP and root-filled pillows were characterized and weighed at 29 DAP. Harvest measurements were similar to those described above except that microbial analyses were not performed. Also, plant height and qualitative status were noted but leaf area measurements were not performed. Photographs and measurements of chlorophyll, fresh mass, dry mass, pillow status, and pillow mass were all taken as previously described.

Data are presented on a per pillow basis. Data were analyzed using GraphPad Prism^® software with reference to Motulsky, 2003. Analysis of variance (ANOVA) was performed on data for chlorophyll, leaf area, height, fresh mass, dry mass, and biomass partitioning (radish). Data are presented as true means. Mean separation was performed by Bonferroni post-tests for two-way ANOVA.

For the broad survey of thirteen species in two types of media, media was not found to be a significant factor in any of the parameters measured across species. Across media, differences between species were found to account for large percentages of the variation in chlorophyll content (59%), leaf area (78%), fresh mass (48%) and dry mass (80%). Greek oregano seeds failed to germinate in pillows with either of the media used, so only twelve species were analyzed, and due to the diversity of chlorophyll levels, size, and growth rate of these species, no general response to media effects can be assessed. Media response was observed in some species, with, for example, delayed growth in a particular type of media, however small sample sizes limit conclusions. Figure 2 shows the dry mass per pillow of the twelve species tested in two media types.

Figure 2.

Cultivars with favorable growth in pillows include ‘Tokyo Bekana’ Chinese cabbage, ‘Sugar Pod II’ snow pea, mizuna, ‘Bright Lights’ Swiss chard, ‘Spicy Globe’ basil and the lettuce varieties ‘Outredgeous’ and ‘Sierra’, based on biomass accumulation and observations of plant health.

Two cultivars of lettuce, ‘Outredgeous’ and ‘Sierra’, were tested for growth in five types of media. Both grew well in pillows and showed no significant variation in leaf area due to different media types. ‘Sierra’ lettuce showed a chlorophyll response to media, with plants growing in perlite: vermiculite having significantly higher SPAD readings than plants growing in arcillite (P<0.05), but no chlorophyll response was observed in ‘Outredgeous’. Fresh and dry mass were significantly different depending on media type. Plants growing in mixtures of Fafard #2 and arcillite had greater mass than those grown in other media, as shown in Figure 3, and in both cultivars, plants grown in 7:3 and 1:1 mixtures of Fafard #2: arcillite showed significantly greater fresh (not shown) and dry mass accumulation than those grown in the 1:1 mixture of perlite: vermiculite. Trends for the observed percentage of root fill in the pillows mirrored those for dry mass (data not shown), with percentages varying from half to full in Fafard #2 and perlite: vermiculite, full in arcillite, and the Fafard: arcillite blends being mostly root-bound.

Figure 3.

Currently the limit for aerobic bacteria on a non-thermostabilized food item is < 2 x 10⁴ colony forming units (CFU) per gram (Perchonok and Douglas, 2012). Studies done to determine the microbial numbers in field and retail samples of a variety of leafy green produce items report total aerobic bacteria levels in the range of 10⁴ to 10⁷ CFU/gram (Johnston et al., 2005; Ruiz et al., 1987). Our analysis of twelve cultivars of leafy greens and herbs indicates microbial densities (2.2 x 10³ to 1.8 x 10⁶) consistent with these studies with the exception of basils, Chinese cabbage, and lettuces grown in arcillite, in which bacterial counts were less than 10⁴/gram. Higher numbers of aerobic bacteria on produce are not necessarily correlated with decreased shelf life, quality or an indication of pathogen contamination (Johnston et al., 2005).

For the cultivars grown in Fafard #2, counts ranged from 1.9 x 10³ on snow pea to the highest bacterial count, 1.8 X 10⁶, on mizuna. Microbial analysis was also performed on one pillow of each ‘Sierra’ and ‘Outredgeous’ lettuce grown on three media types: Fafard #2, arcillite, and 1:1 perlite: vermiculite. Lettuce plants grown in arcillite exhibited aerobic plate counts approximately 1-2 orders of magnitude lower than those grown in perlite: vermiculite and Fafard #2. This information is important in the selection of potential food crops and growth media when considering microbiological standards, however further analysis with greater numbers of replicates needs to be performed, and the more-productive media blends were not tested. Additionally, seeds were not sterilized prior to these tests, and sterilization might lead to reduced microbial levels on produce. Many of the microorganisms detected on the leaves are likely brought in through air flow and human tending, however, so produce sanitation using a sanitizing agent safe for spaceflight is one option to reduce microbial loads in crops and conditions where they exceed allowances.

Several of the highly productive species observed in trial 1 were selected for further study in trial 2. Radish, a crop well-studied for use in space (Goins et al., 2003; Richards et al., 2006) was added for comparison. Due to the media response seen among lettuce cultivars in trial 1, the perlite: vermiculite blend was dropped from consideration and only four media types were examined: Fafard #2, arcillite (1-2 mm), and 7:3 and 1:1 mixtures of the two media. Of the parameters measured, average plant height showed no media dependency (data not shown). Plants grown in pillows reached the same average height at harvest regardless of growth media, indicating that height in these species may be predetermined and not responsive to root zone environments that affect other aspects of plant growth.

Interestingly, qualitative observations of root fill in the pillows also were species-dependent with only slight media differences (data not shown). Snow pea pillows were root bound in all media types, while radish pillows were less than half-full of fibrous roots in all media. Other species were intermediate, with Swiss chard pillows full to root bound, and Chinese cabbage and lettuce pillows half to full. These three species generally showed an increase in roots with increasing levels of arcillite (data not shown).

Chlorophyll content was affected by the type of species and to a lesser extent the type of media. Figure 4 shows that the SPAD readings of radish, snow pea, and chard were not different from one another and averaged in the 40-50 range. Chinese cabbage and lettuce both had lower chlorophyll contents (SPAD~15-30) and were not different from each other regardless of media (Fig. 4). Some media effects were observed, but generally chlorophyll contents (SPAD data) were much more species dependent than media dependent, with species accounting for 80% of the observed variation and media only accounting for 3%. This response is consistent with sufficient macro/micronutrients being available from the slow release fertilizer in the media.

Figure 4.

In contrast to the slight impact of media type on chlorophyll levels, fresh mass was dependent on both media type and species, with species accounting for 33% of the observed variation and media accounting for 28%. Interaction between species and media was not significant. Radish had the greatest fresh mass of all species tested, as seen in Figure 5. For media, Fafard#2 consistently produced plants of all species with the lowest fresh mass, and the Fafard #2: arcillite mixtures tended to produce larger plants, likely related to the balance between moisture holding capacity and aeration of the media.

Figure 5.

Dry mass data displayed an interaction between species and media that was not observed in fresh mass or other parameters, documented in Table 1. Swiss chard, lettuce and snow pea showed no significant response to media and produced the same dry mass as each other. Chinese cabbage plants however, produced a significantly greater dry mass in 1:1 Fafard #2: arcillite than in any of the media tested (P<0.05). The dry mass of Chinese cabbage produced in 1:1 Fafard #2: arcillite was significantly greater than any other species in the same media (P<0.01). Also, Chinese cabbage produced significantly more dry mass than lettuce in every media tested (P<0.05), however there were no differences between the two in fresh mass (see Fig. 5) suggesting a high quantity of water in lettuce compared to the cabbage. Radish plants produced significantly more dry mass in 7:3 Fafard #2: arcillite than in any of the other media types (P<0.05), and since radish showed excellent growth in the 7:3 mixture, a significantly greater dry mass was produced in this media by radish than by either Swiss chard, lettuce or snow pea (P<0.01), and a slightly greater dry mass than Chinese cabbage (Table 1).

General trends indicate that Chinese cabbage and radish dry mass accumulation responds to media more than other species. Although similar in fresh mass, Chinese cabbage generally produces the greatest dry mass, while lettuce the least of any of the crops tested. Radish produced significantly greater dry mass than most other species in the 7:3 mixture, however, these data are total biomass. When only the tap root is examined this value falls to 1.95 g dry mass on average. Table 2 provides an estimate of the maximum growth rate of each of these species. These edible growth rates are comparable to those for lettuce and soybean observed by Wheeler et al., 2003, though they are lower than for other staples such as potato.

Interestingly, although total fresh mass of radish demonstrated a significant influence of media, individually, there was no significant difference between shoots grown in different media types or roots grown in different media types. Figure 6 shows that dry mass, however, was significantly different, with plants grown in 7:3 Fafard: arcillite having significantly greater shoot mass than arcillite- or 1:1-grown plants and significantly greater root mass than Fafard #2-grown plants.

Figure 6.

When the percentage of biomass in the root and shoot are calculated we observe increasing partitioning into the edible root with increasing concentration of arcillite and decreasing percentage of Fafard #2 in the media, as shown in Figure 7, and the interaction between media and the location of partitioning is highly significant (P<0.0001).

Figure 7.

The observed differences in response to media may be due to a variety of media characteristics including moisture, aeration, and cation and anion exchange capacity, however since water is not leaching through the pillows it seems likely that water and aeration may be playing the largest role. The arcillite, a porous clay, has a lower water holding capacity (~50% by weight) than the peat-based Fafard #2 medium which expands when wet and can hold ~60% moisture by weight. The physical and chemical parameters of the media can lead to differences in root growth and nutrient availability, and several of these parameters are gravity-dependent. Future testing will further assess these challenges; however initial results using the pillow concept indicate that this technique may be a good approach to growth of salad and herb crops for the Veggie hardware.