Introducing short roots in a desert perennial: anatomy and spatiotemporal foraging responses to increased precipitation


  • Roberto Salguero-Gómez,

    1. The University of Pennsylvania, Department of Biology; Leidy Laboratories 321, 433 South University Avenue, Philadelphia, PA-19104-6018, USA; Present address: Max Planck Institute for Demographic Research, Konrad-Zuze-strasse 1. 18057 Rostock, Germany
    Search for more papers by this author
  • Brenda B. Casper

    1. The University of Pennsylvania, Department of Biology; Leidy Laboratories 321, 433 South University Avenue, Philadelphia, PA-19104-6018, USA; Present address: Max Planck Institute for Demographic Research, Konrad-Zuze-strasse 1. 18057 Rostock, Germany
    Search for more papers by this author

Author for correspondence:
Roberto Salguero-Gómez
Tel: +1 215 898 8608


  • The desert flora possesses diverse root architectures that result in fast growth in response to precipitation. We introduce the short root, a previously undescribed second-order root in the aridland chamaephyte Cryptantha flava, and explore fine root production.
  • We describe the short root anatomy and associated fine roots, correlate standing fine root crop with soil moisture, and explore the architectural level – the short root, third-order lateral roots, or the whole root system – at which fine roots are induced by watering and the amount of water required.
  • We show that short roots are borne at intervals on lateral roots and produce fine roots at their tips; new fine roots are white and have root hairs, while brown and black fine roots are apparently dead; and fine root production is triggered at the level of lateral roots and with relatively low precipitation (≤ 2 cm).
  • Short roots are suberized and thus are probably not capable of water uptake themselves, but serve as initiation sites for fine roots that grow rapidly in response to rainfall. Thus, C. flava should be a beneficiary of projected precipitation increases in habitats where rainfall is pulsed.


Life in arid ecosystems is primarily limited by the overall low, yet temporarily and spatially variable water availability. Water modulates microbial decomposition and nutrient cycling (Steinberger et al., 1995), facilitates nutrient uptake and transport (Holbrook & Zwieniecki, 2005), and enables photosynthesis. In cold deserts, water is made available by the melting of snow accumulated during the winter and via precipitation pulses during the growing season, which vary greatly in intensity and frequency (Dobrowolski et al., 1990). In addition, the spatial variation of resources – including soil moisture – in deserts may be as large within the rooting area of a plant as the variation in the entire community (Jackson & Caldwell, 1993).

Desert plant species possess a wide array of strategies to cope with such temporal and spatial water variability. Some examples include tolerance to long drought periods via extensive root systems that tap into a large soil volume (Mooney et al., 1980; Canadell et al., 1996), roots that store water (Graham & Nobel, 1999), and spatiotemporal partitioning of the soil moisture by precipitation intensity via differential root depth and seasonal root activation (Lin et al., 1996; Schwinning et al., 2003; Ogle & Reynolds, 2004). Furthermore, desert plants are extremely phenotypically plastic and can quickly produce new roots when it rains (Nobel & Sanderson, 1984; Jackson & Caldwell, 1989) and abscise them when the soil dries out (North et al., 1993).

General circulation models project a 25–50% increase in summer precipitation in the coming decades in the arid southwest of the United States (Arritt et al., 2000; Easterling et al., 2000). However, regional climatic models for the Great Basin desert currently differ on whether this increase in precipitation will take place in the form of more frequent, smaller pulses (Sala & Lauenroth, 1982; Field et al., 1999), or fewer, more intense pulses (Groisman et al., 1999; TNAS, 2000). These different pulse intensities may vary in infiltration and runoff (Wainwright et al., 1999) and thus may have consequences for the physiology, demography and community composition of the native flora (Schwinning et al., 2003; Huxman et al., 2004).

Although much work has been devoted to the study of above-ground responses of desert plants to precipitation (Lin et al., 1996; BassiriRad et al., 1999; Gebauer & Ehleringer, 2000; Huxman et al., 2004; Schwinning et al., 2004), we still lack a full understanding of their below-ground responses. This knowledge is critical in the light of climate change because root responses typically precede above-ground changes (Fernandez & Caldwell, 1975), because a large proportion of desert plant biomass is below ground (Canadell et al., 1996), and because the desert flora often operates close to hydraulic failure under the current climatic conditions (Davis et al., 2002; Ackerly, 2004).

In the present study, we explore below-ground responses of the aridland species Cryptantha flava to spatial and temporal variation in precipitation intensity. Previous studies have shown that C. flava can respond to late-summer rainfall by creating new sets of leaves (Casper et al., 2001), that the carbon thus assimilated carries over to the following spring (Casper et al., 2005), and that naturally occurring large fall precipitation augments seedling establishment and contributes to population growth (Lucas et al., 2008). Here, we first offer an anatomical description of a unique root type found in C. flava, the short root and its associated fine roots. We then examine the abundance of fine roots in relation to natural soil water conditions throughout the growing season and under an experimentally prolonged growing season, where we simulated pulses of various intensities to explore thresholds of fine root production. Finally we examine the ability of C. flava to forage under spatial heterogeneity by watering different soil sectors around a plant and recording its root responses.

Materials and Methods

Study species and field site

Cryptantha flava (A. Nels.) Payson (Boraginaceae) is a chamaephyte (herbaceous-subshrub; Raunkiaer, 1934) species common throughout the Colorado Plateau in the USA (McLaughlin, 1986). Leaves are normally present from late March through early August unless extended by late season rains (Casper et al., 2001). Flowering occurs during May–June, and seeds mature by mid-July. Most seeds germinate either in September–October or at the beginning of the following growing season (B. B. Casper, pers. obs.). Leaf rosettes are grouped into modules, each connected to a different branch of the caudex (Fig. 1a), the underground stem. There is a single taproot, extending at least 1.5 m deep, from which several lateral roots branch off at depths of 20–40 cm. Lateral roots spread horizontally up to c. 0.9 m in juveniles and c. 1 m in adults (Peek & Forseth, 2005) before turning downwards at the ends. Fine roots are produced in clusters along the lateral roots.

Figure 1.

Architecture of an adult of Cryptantha flava and its root system. (a) Groups of leaf rosettes develop from different shoots (modules), but they all branch off from the same caudex, the central, underground stem. The root system consists of a deep tap root and a discrete number of lateral roots that spread laterally. (b) Lateral roots (LR) have short roots (SR), from which fine roots of different colorations develop: white (WFR, white arrow), brown (BrFR, orange arrow) and black (BlFR, black arrow) fine roots. (c) Portion of lateral root containing white, brown and black fine roots. Note that only the white fine root possesses root hairs (RH), while the brown and black fine roots are thin and seemingly not functional. (d) Unstained 30-μm cross-section of a white fine root with its epidermal cells (Ep), cortex (CX), endodermal cells (En), and the stele (S), which contains the xylem (X) and phloem vessels (P). (e, f) Toluidine-blue stained 8 μm longitudinal section and cross-section, respectively, of the vasculature of the lateral root and its connection with the short root from the pericycle (PC) of the lateral root. (g) Stained 5 μm longitudinal section of a short root, showcasing the development of a white fine root from the apex of the short root, and the connections of brown and black fine roots at the base of the short root. The short root is covered by suberin (SB) and hosts chambers from which fine roots develop (SRC).

The hydraulic design of C. flava changes with ontogeny. In juveniles, below-ground resources gathered by all lateral roots and the taproot are shared throughout the entire individual, but as plants grow they become internally fragmented, resulting in several ‘integrated hydraulic units’ (sensuSchenk, 1999); soil resources taken up by a specific lateral root travel to a specific rosette module (Salguero-Gómez & Casper, 2011). The roots of adults also differ in having a thicker covering of suberin, a hydrophobic protein, which likely reduces water loss during droughts (R. Salguero-Gómez, pers. obs.).

Experiments were conducted and material for histological examination collected at a site managed by the Bureau of Land Management near the Redfleet State Park (40°30′N, 109°22′30″W, 1730 m above sea level (asl), northeast Utah, USA). The vegetation is dominated by woody species (Juniperus osteosperma, Artemisia tridentata and Chrysothamnus nauseosus); C. flava is the dominant chamaephyte. The soil is an Aridisol, and C. flava thrives where a discontinuous, thin hardpan (10–15 cm deep) is not present.

Root histology

For histological examination of fine root clusters, we collected segments of lateral roots during summer 2007 and fixed them in situ with FPA (Paraformaldehyde: 30% ethanol, 5% propionic acid, 5% formaldehyde, 50% distilled H2O, with an additional 10% glycerol to soften tissue for later sectioning). Samples were dehydrated in a graded ethanol series, embedded in histological resin (Technovit 7100; Heraeus Kulzer, Frankfurt, Germany), sectioned at 10 μm using a rotary microtome with a tungsten blade (Reichert-Jung, D-profile; Leica, Bannockburn, IL, USA), and stained with toluidine blue (0.5%, v/w). Longitudinal sections and cross-sections were photographed (Olympus BX51; Olympus, Allentown, NJ, USA) at ×10 and ×40 magnification.

Climatic data

To compare the timing and amount of our experimental watering to natural patterns of precipitation, we obtained records (1931–2008) from a permanent meteorological station, ‘Maeser 9 NW’, located 18 km away in NE Utah, USA, at 1950 m asl (Western Regional Climate Center, Because in cold deserts the build-up of soil moisture from snow can be important for the following growing season (Ogle & Reynolds, 2004), we compared annual precipitation from September of the preceding year to August of the studied year (2006–07 or 2007–08) with the long-term mean (1931–2008). We also estimated from the long-term climatic data the probability that a precipitation event equal in intensity to each of our experimental watering treatments would occur within each calendar month.

Experimental design

In order to help understand the ecology of fine roots, we measured the standing number of fine roots under natural amounts of soil moisture and also fine root production in response to experimental watering treatments. We were interested both in the quantity of water needed to induce fine roots and in whether induction is controlled at the level of the whole plant, the entire lateral root or the specific site of a fine root cluster.

Standing fine root crop during the growing season  We collected lateral root segments throughout the growing season and compared the number of live fine roots with soil moisture content. We removed c. 5-cm-long segments of lateral roots from eight arbitrarily chosen individuals on 13 different days at c. 7 d intervals beginning the last week of May until the last week of August, in both 2007 and 2008. Roots were obtained at c. 25 cm depth and 20–30 cm away from the caudex and preserved immediately in 50% ethanol. We never re-sampled individuals. We quantified the standing fine root crop as the total length of white, apparently live and functional, fine roots (Fig. 1) on each lateral root segment, normalized by the actual length of the sampled lateral root (mm fine root mm−1 lateral root). Each fine root was measured to the closest 0.1 mm with a caliper under a dissecting microscope (Olympus MV Plapo 2XC). We also collected soil surrounding the sampled lateral root segment to calculate the gravimetric soil water content (SWC; Pearcy et al., 1989). We counted the total number of leaf rosettes, which correlates strongly with above-ground biomass (Salguero-Gómez & Casper, 2011), and classified individuals as juveniles (nonflowering, < 25 rosette individuals with no evidence of having shrunk from a larger size) or adults (≥ 25 rosettes) to test whether ontogenetic stage and SWC affected standing fine root crop using a two-way ANOVA.

We tested whether precipitation affects fine root production by carrying out three analyses (ANCOVA) for each year with plant size as covariate, standing fine root crop as the response variable, and either SWC, monthly precipitation registered during the calendar month of root collection, or monthly precipitation registered during the calendar month immediately preceding root collection date as the explanatory variable. The data met normality assumptions. We then fitted linear and polynomial regressions to describe changes in standing fine root and SWC over the 2007 and 2008 growing seasons separately and used adjusted R2 values to determine the best model.

Moisture threshold for producing fine root growth  We examined fine root production in response to simulated pulses of precipitation in August, both in 2007 and 2008. We chose three locations, 100 m apart, and supplied no water (0 cm), or a one-time pulse of precipitation of 2, 4.5 or 7 cm evenly distributed over the 1 m radius of a circle centered on each individual (hereafter called the homogeneous watering treatment). Long-term data show that these amounts occur naturally with probabilities of 0.7–0.015 in August (Supporting Information, Fig. S1). To ensure accurate deliveries of water pulses, we placed small rain gauges within the watered area. Each arbitrarily selected plant (ntotal = 72) was watered once between 17:00 and 18:00 h. All watering took place over 3 d during the second week of August, when natural monsoonal precipitation might occur. No natural precipitation occurred during or 2 wk before water in 2008, but a light rain (< 0.3 cm) occurred during the first week of August 2007. Before watering, for each individual, we counted the number of rosettes and collected c. 100 g of soil 1.5 m south at 25 cm depth between 16:00 and 18:00 h in order to obtain pretreatment gravimetric SWC (SWCprewatering).

Between 16:00 and 18:00 h on the seventh day after watering an individual, we collected one 5-cm lateral root segment from its east side and one from its west side at a depth of c. 25 cm and 20–30 cm from the caudex. We counted white fine roots and measured the standing fine root crop as previously described. At the time of root collection, we also collected c. 100 g of soil neighboring the lateral root in order to measure post-treatment SWC (SWCpostwatering). We calculated the net change in SWC caused by our watering treatment as ΔSWC = SWCpostwatering– SWCprewatering.

We used linear regressions to examine the effect of pulse intensity on standing fine root crop and on ΔSWC. We tested significant differences among pulse intensities in standing fine root crop and in ΔSWC using the Tukey–Kramer HSD tests (Sokal & Rohlf, 1995). As the date of sample collection, plant size (regression analyses), and whether the root segment came from the east or west side of the plant (paired t-test) did not significantly affect either fine root crop or ΔSWC, we excluded them from posterior analyses.

Degree of independence of fine root growth  Because we were interested in a plant’s response to soil water heterogeneity, given the species’ fragmented architecture (Salguero-Gómez & Casper, 2011), we supplied pulses of precipitation to different portions of the root system to determine the spatial scale at which water induces fine root production.

First, we tested whether fine roots are produced in response to water applied to a cluster of fine roots. We arbitrarily chose 16 individuals in late May 2007, carefully excavated one live lateral root, and recorded the number and measured the cumulative length of live fine roots within a single cluster using a caliper. For eight individuals, we carefully placed beneath the lateral root a 30 ml plastic cup filled with water and used a sponge to wick water to that single cluster of fine roots. One end of the sponge was placed in the water and the other wrapped around the lateral root and fine root cluster and covered with plastic wrap to prevent conduction of water to neighboring portions of the root. In the remaining eight individuals, we tagged a targeted cluster of fine roots with tape on the lateral root and measured the number and cumulative length of white fine roots. In both cases, watered and unwatered, we replaced the soil and re-measured the number and cumulative length of white fine roots within the targeted fine root clusters 7 d later. No natural precipitation occurred during this experiment. We used repeated-measures ANOVA (Sokal & Rohlf, 1995), with watering treatment and time of measurement as explanatory variables, to examine cumulative live fine root length at days 0 and 7. Because six of the 16 targeted fine root clusters had no live fine roots, data were not normally distributed. We excluded these from the analyses, but we report their responses here.

Failing to induce fine roots at the level of the fine root cluster, we then tested whether fine roots are induced along lateral roots when the soil around them is watered. We arbitrarily selected individuals in each location used for the homogeneous watering treatment and watered a 60° sector of a 1-m-radius circle centered on each plant (hereafter called the heterogeneous watering treatment). We simulated precipitation pulses of 2, 4.5 and 7 cm on 15, 15 and 10 individuals per location and year, respectively (ntotal = 240 plants). Watering took place at the same time as the homogeneous watering treatment, and 7 d later we collected two lateral root segments per individual, one from the watered sector and the other 180° away in the unwatered sector. We counted the number of live fine roots on each sample and measured their standing fine root crops. We collected soil 1.5 m south of each individual at the onset of the experiment and at the time of root collection immediately adjacent to each sampled lateral root in order to calculate ΔSWC. For each year, we carried out two-way ANOVAs with pulse intensity and sector (watered or unwatered, paired by individual as a random variable) and as explanatory variables and standing fine root crop or ΔSWC as response variables. Location, plant size and date of water application were not significant effects.

To determine whether fine root growth on lateral roots in watered sectors is the same as when the whole root system is watered, we compared standing fine root crop and ΔSWC between the heterogeneous and homogeneous watering treatments each year. First, we made comparisons between the 0 cm pulse of the homogeneous watering treatment and the unwatered sector of the heterogeneous treatment using t-tests. Next, we made comparisons between the homogeneous watering treatment and the watered sectors of the heterogeneous watering treatment using two-way ANOVAs with treatment and pulse intensity (2, 4.5 and 7 cm) as explanatory variables.

Finally, we determined whether the standing fine root crop induced by each of our watering treatments at the end of the growing season differed from the standing fine root crop under a comparable SWC during the growing season. To do so, we tested whether the slope of the relationship between SWC and standing fine root crop differed for natural amounts of precipitation and our experimentally manipulated pulses. We used a two-way ANOVA with standing fine root crop as the response variable and SWC and pulse intensity (natural amounts or pulse intensities of 2, 4.5, or 7 cm) as explanatory variables. Because in both years, the SWC × pulse intensity interaction was not significant, we carried out ANCOVAs with SWC as main effect and pulse intensity as the covariate to compare the relationship between SWC and fine root crop among pulse intensities. We used JMP 8.0 (SAS Institute, Cary, NC, USA) for all statistical analyses.


Histology of short roots

In C. flava, fine roots, first-order roots are produced exclusively on second-order short branch roots, which in turn are borne on the third-order lateral roots. We call these second-order roots ‘short roots’ because their lack of elongation and their production of fine roots are analogous in function to nonelongating short shoots in many woody species (Fig. 1e,g) (the structure short root described here is not to be confounded with the SHORT ROOT gene (SHR), which controls asymmetric cell division and formation of root endodermis and cortex (Helariutta et al., 2000)). Our histological investigations reveal that short roots develop from the pericycle of the lateral root, rupturing the cortex as they emerge (Fig. 1f,g), and they connect the vasculature of the lateral and fine roots (Fig. 1g). Most of the short root is covered by the waxy protein suberin (Fig. 1f). Short roots average 2.2 ± 0.4 (SE) mm in width and 2.7 ± 0.3 (SE) mm in length.

Fine roots exist in three color-classes: white – the only ones apparently alive –, brown and black. These differ in their cellular integrity and spatial positioning on the short root: white fine roots are found at the apex, brown fine roots lie immediately posterior to the white ones, and black fine roots occur at the base of the short root (Fig. 1e,g). Unicellular root hairs are only present on white fine roots (Fig. 1c,d), which have a one- to two-cell epidermis, a cortex two to three cells thick, and a typical one-celled endodermis and stele (Fig. 1d). Neither subcellular components nor vascular tissues are identifiable in brown and black fine roots, which are significantly more fragile and c. threefold thinner than white fine roots, and withered. At any one time, a short root typically supports either none or just one white fine root and one to six brown or black fine roots in total. Short roots contain a significant proportion of cortical cells (Fig. 1f,g), considerably more than the lateral root.

Climate patterns

The long-term (1930–2008) annual mean precipitation at the Maeser meteorological station is 25.11 cm, but the amount of precipitation has increased significantly during the past 78 yr (annual precipitation: = 43.76, df = 76, < 0.001), with the greatest monthly increase in August (Table S1). Annual precipitation was below the mean in 2006–07 (28.11 cm) and above the mean in 2007–08 (36.34 cm). In 2006–07, winter precipitation was above the mean but the growing season was drier. Precipitation in 2007–08 was more evenly distributed throughout the year (Fig. 2a). The probability of a precipitation event comparable to intensities used in our experiment occurring within the span of a week varies monthly. The baseline probability for no precipitation is quite constant throughout the year (pulse intensity = 0 cm; average probability = 0.67 ± 0.02); large precipitation events are more likely in May, August and October (Fig. S1).

Figure 2.

Precipitation and root responses during the study. (a) Monthly mean precipitation in relation to the 1930–2009 record (light gray, above mean; dark gray, below mean). (b) Gravimetric soil water content (closed circles) and amount of fine roots (open circles) of Cryptantha flava as a function of the collection time during the growing seasons of 2007 and 2008. Error bars are ±SE. Black and gray lines indicate polynomial regressions fitted to the soil water content and standing fine root crop, respectively.

Standing fine root crop relates to soil water content

Under natural precipitation, the standing fine root crop of C. flava was significantly lower during the 2007 growing season (0.17 ± 0.02; inline image ± SE) than in 2008 (0.24 ± 0.02; = 6.32, df = 377, = 0.01). The standing fine root crop of an individual was positively correlated with local SWC for both years (2007: = 5.81, df = 190, = 0.01; 2008: = 9.27, df = 185, < 0.001; Fig. 2b) but not with ontogenetic stage (juvenile or adult) or its interaction with SWC. Furthermore, soil sample SWC and standing fine root crop were both positively correlated with the amount of precipitation received during the calendar month of each sample collection (SWC, 2007: = 7.30, df = 190, < 0.001; 2008: = 5.92, df = 185, < 0.001; fine root crop, 2007: = 1.89, df = 190, = 0.06; 2008: = 4.02, df = 185, < 0.001) but not with precipitation in the previous month. In both years, SWC and standing fine root crop decreased during the growing season, but then increased in mid-August, coinciding with the arrival of monsoonal rainfall events (Fig. 2b). Such responses were satisfactorily described by second-degree polynomial regressions; for the standing fine root crop (SFRC) in 2007, SFRC = 0.58–2.45 × 10−3 × d + 3.61 × 10−5× (d − 189.19)2 (adjusted R2 = 0.75, < 0.001) and in 2008, SFRC =  1.31–6.25 × 10−3 × d + 9.85 × 10−5(d − 186.08)2 (adjusted R2 = 0.83, < 0.001); and for the SWC in 2007, SWC = 0.02–7.25 × 10−5 × d + 3.27 × 10−6 (d − 188.85)2 (adjusted R2 = 0.51, = 0.011), and in 2008, SWC = 0.11–5.31 × 10−4 ×d + 1.07 × 10−5 (− 186.07)2 (adjusted R2 = 0.94, <0.001), where = day of year.

Threshold for fine root growth

The standing fine root crop in C. flava increased with the application of water of any intensity (2, 4.5 and 7 cm) in both years (Fig. 3a). The increase was driven both by the production of new white fine roots and the elongation of existing fine roots; fine roots were present on 14% (2007) and 35% (2008) of the sampled short roots in all watering treatments combined (2, 4.5 and 7 cm), while white fine roots were present only in 4% (2007) and 17% (2008) in controls (0 cm). The average length of white fine roots in the watering treatments was 11.41 ± 2.75 mm (2007) and 9.65 ± 2.32 mm (2008) and, in control plants, was 0.88 ± 0.19 mm (2007) and 3.97 ± 2.12 mm (2008).

Figure 3.

Effect of the homogeneous watering treatment with different pulse intensities on the fine roots of Cryptantha flava (a) and soil water content (b) in August 2007 and 2008. Bars with the same uppercase letters did not differ significantly in Tukey’s HSD adjusted comparisons for each panel separately. Error bars are ±SE.

Although ΔSWC was positively and linearly affected by the pulse intensities used (2007: = 3.95, df = 62, < 0.001; 2008: = 2.81, df = 56, < 0.001; Fig. 3b), the standing fine root crop did not exhibit a monotonic increase as a function of pulse intensity; there was a significant increase in live (white) fine root abundance between 0 and 2 cm of precipitation (< 0.005 for both years), but the standing fine root crop did not differ among pulses of 2, 4.5 and 7 cm in either year (Fig. 3a).

Degree of independence in fine root growth

Fine root number and cumulative length had not increased by 7 d after the application of water to individual fine root clusters (short roots) of C. flava (Table S2). Furthermore, short roots that did not support live fine roots before watering them individually had not developed new live fine roots 7 d later. By contrast, in the heterogeneous watering experiment, the standing fine root crop was significantly higher in the root sector that received water than in the unwatered sector, regardless of pulse intensity (Fig. 4, Table 1).

Figure 4.

Effect of the heterogeneous watering treatment, where one sector of the Cryptantha flava plant was watered with different pulse intensities (closed bars) but another was not (open bars), on the fine roots (a) and soil water content (b) in August 2007 and 2008. Error bars are ±SE. Bars with the same upper-case letters did not differ significantly in Tukey’s HSD adjusted comparisons for each panel separately.

Table 1.   Summary of the two-way ANOVA for the responses of the fine root growth of Cryptantha flava and net change in soil water content in the heterogeneous watering treatment, where one sector of the plant was watered with pulses of different intensity (2, 4.5 and 7 cm) and the remaining sector of the plant was unwatered (0 cm)
EffectsStanding fine root cropNet change soil water content
dfSum of squaresF-ratioP-valuedfSum of squaresF-ratioP-value
  1. *, < 0.05; **, < 0.005; ***, < 0.001.

 Sector11.1023.46< 0.001***10.0019.510.002**
 Sector × pulse10.0020.040.8310.0102.120.15
 Sector × pulse10.0100.170.6910.0000.120.73

The standing fine root crop for sectors that did not receive water in the heterogeneous treatment (0.013 ± 0.003 mm mm−1 in 2007, 0.072 ± 0.0337 mm mm−1 in 2008) was not different from roots in the homogeneous watering experiment where no water was delivered to the plant at all (0.004 ± 0.007 mm mm−1 in 2007, 0.007 ± 0.087 mm mm−1 in 2008; Table 2a). Likewise, the standing fine root crop for plants receiving water in the homogeneous treatment and for watered sectors of the heterogeneous watering treatment did not differ for any pulse intensity (Table 2b). Across all watering intensities, the standing fine root crop in watered sectors of the heterogeneous treatment was 0.17 ± 0.03 mm mm−1 in 2007 and 0.22 ± 0.03 mm mm−1 in 2008 and, in watered plants of the homogeneous treatment, 0.17 ± 0.04 mm mm−1 in 2007 and 0.22 ± 0.04 mm mm−1 in 2008.

Table 2.   (a) Summary of t-tests for fine root production of Cryptantha flava and net change in soil water content in soils that did not receive water, from the 0 cm pulse in the homogeneous watering treatment and from the nonwatered sectors of the heterogeneous watering treatment. (b) Summary of two-way ANOVA for the responses of the fine root growth and net change in soil water content where plants were either watered in one sector or in its entire surrounding surface with watering pulses of various intensities (2, 4.5 and 7 cm)
EffectStanding fine root cropNet change soil water content
  1. ***, < 0.001.

(a) Unwatered soil
(b) Watered soil
  Pulse10.0940.760126.113< 0.001***
  Treatment × pulse10.18850.66510.3140.576
  Treatment × pulse10.0940.76010.9310.336

No detectable amount of water moved from watered to unwatered soil sectors in the heterogeneous watering experiment. SWC in the unwatered soil sector of the heterogeneous watering experiment remained constant regardless of pulse intensity in 2007 and decreased slightly in 2008 (Fig. 4b, Table 1). Furthermore, the ΔSWC of the unwatered soil sectors of the heterogeneous watering treatment and the soils of 0 cm intensity in the homogeneous watering treatment did not differ (Table 2a).

Retrospective comparison on the production of fine roots

The relationship between SWC and standing fine root crop (Fig. S2) was the same in the watering treatments (2, 4.5 and 7 cm) at the end of the growing season as under natural amounts of precipitation throughout the growing season. While the overall two-way ANOVA model comparing root standing crop among water treatments and natural amounts of precipitation was significant (2007: F7, 229 = 8.26, = 0.03; 2008: F7, 226 = 12.41, < 0.001), neither pulse intensity alone nor its interaction with SWC had a significant effect. The ANCOVA tests with SWC as a main effect and pulse intensity as covariate revealed a significant effect of the former on standing fine root crop (2007: = 0.02; 2008: < 0.001), but pulse intensity, including natural amounts of precipitation, was not significant.


An undescribed root type: the short root

To our knowledge the root type introduced in this study, the short root, has not been previously described. This structure is at least present in a congener (Cryptantha flavoculata; R. Salguero-Gómez, pers. obs.), and there exists photographic evidence for another desert perennial in a different family (Franseria deltoidea, Compositae; Cannon, 1911; p. 73), although those roots were not examined histologically. The name ‘short root’ was chosen because of its analogous design to short shoots in taxa such as Ginkgo biloba (Gunckel & Wetmore, 1946).

We suggest that the short root anatomy in C. flava is advantageous in a water-limited environment for at least three reasons. First, it enables the rapid production of fine roots in response to sufficient precipitation just as short shoots of some desert shrubs quickly produce new leaves (Edwards & Diaz, 2006). In both short shoots and short roots, little growth of the apical meristem is necessary in order to generate lower rank structures (leaves or fine roots). We found fine roots 2 d after a simulated rainfall of 4.5 cm (R. Salguero-Gómez, unpublished). Second, the clustering of fine root production on short roots, and thus in discrete locations, minimizes the exposure of more permeable root surface to dry soils because the remainder of the lateral root is suberized, although the taproot of C. flava is less suberized with increasing depth. Third, the short root serves as a short connector pathway between fine roots and lateral roots, which with their high proportion of conducting tissue are suited for long distance transport. However, lateral roots of C. flava do have some cortex, which is lacking in second-order roots of trees (Guo et al., 2008).

Because of their rapid response, fine roots in C. flava perform analogously to ‘rectifier roots’ of some succulents in response to water (Nobel & Sanderson, 1984), proteoid roots in response to rich phosphorus and iron microsites (Watt & Evans, 1999), or pad adventitious roots in climbing vines (Shishkova et al., 2007). However, the branching architecture of C. flava differs structurally from these other examples. Fine ‘rain roots’ of succulents are mostly found on thin lateral roots several branching orders removed from the taproot; these lateral roots are not as well suberized (Gibson & Nobel, 1990), and the rain roots themselves are more ephemeral (Snyman, 2006) than the fine roots of C. flava that may continue growing for at least 4 d after watering (R. Salguero-Gómez, pers. obs.). In addition, lateral roots of cacti are usually closer to the soil surface than those of C. flava, and so precipitation pulses of very low intensity can activate their fine root production (Dougherty et al., 1996).

Although we do not have direct evidence, we suspect that increased coloration of fine roots correlates with age, based on the fact that only white roots possess root hairs, the darkest fine roots are located farthest from the short root apex, and dark roots show a general degradation in internal structure. Thus fine root coloration matches a chronological sequence if fine roots are always produced near the apex of the short root. Additionally, mycorrhizal vesicles, which may resist unfavorable environmental conditions (Harrier, 2001), are primarily found in brown and black fine roots, while hyphae and arbuscules are visible in white ones (Fig. S3). While fine roots in other species may become more pigmented with age (Van Rees et al., 1990), the extent to which they lose function is debated; in some species, dead fine roots are still able to absorb water (Eissenstat & Yanai, 1997; Comas et al., 2000). The function of brown and black fine roots in C. flava deserves further attention.

Spatial heterogeneity and resource foraging

While fine roots are confined to short roots, our watering experiments show their production occurs in response to soil moisture cues received at a higher level of root architecture, though still at a level below the whole root system. This ability should position C. flava as an effective forager for spatially heterogeneous soil resources typical of deserts (Forseth et al., 2001; Maestre et al., 2005).

Our results here enable us to link the activity of individual lateral roots in C. flava to the spatial distribution of soil resources within the plant as determined in a previous study. In adult plants, but not in juveniles, the movement of water between lateral roots and clusters of leaf rosettes is highly sectored whereby a single lateral root supplies a particular portion of the canopy (Salguero-Gómez & Casper, 2011), a condition known as hydraulic sectoriality (Orians et al., 2005). While the single lateral root of a juvenile might supply the entire plant with water or even, we hypothesize, allow horizontal hydraulic redistribution through the root system from a wet patch of soil to a dry one (Bauerle et al., 2008), an adult’s lateral root located in a particularly dry microsite and the particular rosette modules to which it is connected could experience severe water stress without the possibility of augmentation by the remainder of the root system. Indeed, following drought, larger individuals of C. flava are more likely to show partial canopy death, or shrinkage, than smaller individuals (Casper, 1996; Lucas et al., 2008). Juveniles are also less reliant on lateral roots and depend more on the taproot than adults (Salguero-Gómez & Casper, 2011). Thus root anatomy and development are linked to whole-plant physiology and, in turn, to demography. A full understanding of fine root development in C. flava in response to water would include information on genetic and developmental mechanisms, including relevant signal transduction pathways.

Temporal variability and climate change

The appearance of new white fine roots in response to watering in August and September suggests that C. flava will exploit end of growing season increases in precipitation projected by some climate change models (Schlesinger & Mitchell, 1987; Lin et al., 1996). Above average precipitation during those months has induced new cohorts of leaf rosettes in C. flava at our field site, prolonging the growing season by over a month (Casper et al., 2001). Lin et al. (1996) suggest that herbaceous species are better equipped to profit from late-season increases in precipitation than woody species because of their shallower root architecture and greater phenotypic plasticity. As a chamaephyte, C. flava has a woody caudex with secondary growth and a deep taproot, but it resembles herbaceous species in its ability to respond rapidly to rainfall events. Our finding some white fine roots throughout the growing season provides morphological confirmation that C. flava is capable of taking up water and other soil resources throughout the summer even though it responds strongly to punctuated rainfall events, particularly at the end of the growing season (Gebauer & Ehleringer, 2000). The advantages of perennating woody tissue together with the ability to exploit shallow, temporally available soil resources may help to explain the commonness of the chamaephyte growth form throughout much of the cold and arid Colorado Plateau (McLaughlin, 1986).

Thresholds and climate change: small vs large pulses

Identifying thresholds of soil moisture responses in different species is critical to understanding how changes in the intensity and frequency of rainfall events, as predicted by climate change models, will impact community composition. Intense pulses can recharge the water table, and thus are expected to benefit deep-rooted woody species, whereas small pulses reach limited soil depths, which is where herbaceous species typically forage (Sala & Lauenroth, 1982). Biological responses may not scale linearly with water availability, as suggested by the ‘pulse-reserve’ paradigm (Noy-Meir, 1973), but exhibit a threshold of activation (Reynolds et al., 2004; Schwinning et al., 2004). Here, we show that C. flava has a relatively low threshold for activation of fine root growth (≤ 2 cm) and that pulses of higher intensity do not result in greater root production. We can think of several explanations for this asymptotic scaling: fine root production may be limited by the plant’s carbon budget or physically limited by the number of potentially active short roots; additional water uptake might take place through the suberized root layer under very high SWC, as reported for Atriplex confertifolia (Caldwell & Camp, 1974); and/or regardless of the permeability of lateral roots, additional fine roots might not further increase the water uptake rate. We can reject the possibility that 2 cm pulses are enough to saturate the soil (Sala & Lauenroth, 1982) because increasing pulse intensities resulted in wetter soils (Fig. 3b). Leaf water potentials and/or the use of isotopically labeled water might reveal whether water uptake increases with pulse size even without additional fine roots.

Final remarks

Our description of short roots in C. flava and their ability to produce fine roots rapidly in response to soil water adds to the documented repertoire of strategies plants use to cope with spatiotemporal resource variability characteristic of arid ecosystems. Our finding that fine root production can be induced at a scale smaller than the whole root system means an individual can utilize spatially heterogeneous soil resources. The continued production of fine roots in response to late season water causes us to predict benefits for this species, as well as others of similar morphology, from increased late season precipitation forecasted for the Great Basin desert by climate change models.


We thank S. Poethig, M. Peek, D. Eissenstat, T. Rost and two anonymous reviewers for improving previous versions of the manuscript; J. Salix and T. Faircloth (Vernal Bureau of Land Management, UT, USA) for logistical support during field work; M. Peek for very useful discussions about the root ecology of Cryptantha flava, and B. Waring and A. Zeng for assistance in the field. Microtomy and microscopy were carried out at S. Poethig’s laboratory (University of Pennsylvania, USA) and F. Ojeda’s lab (Universidad de Cadiz, Spain). We are especially thankful to A. Santos-Alvarez for assistance with resin histological techniques. R.S-G. was supported during fieldwork by the Forrest Shreve Desert Ecology award of the Ecological Society of America, the GIAR from Sigma Xi, and the Binns-Williams funds of the Biology Department of the University of Pennsylvania.