Functional 
Ecology 1992 
6,575-581 
Ecological functions of carbohydrates stored in corms 
of Tipularia discolor (Orchidaceae) 
J. K. ZIMMERMAN and D. F. WHIGHAM* 
Center for Energy and Environment Research, University of Puerto Rico, GPO Box 363682, Sun Juan, Puerto 
Rico 00936 and *Smithsonian Environmental Research Center, PO Box 28, Edgewater, Maryland 21037, USA 
Summary 
1. The  orchid Tipularia discolor possesses two to  five subterranean corms containing 
a high concentration of total non-structural carbohydrates (TNC). W e  investigated 
the  role of corms in the  initiation of new growth following seasonal dormancy, in 
reproduction, and in the  response of plants to  artificial defoliation. 
2. Severance of older corms (representing c. 3O0/0 of total corm biomass) influenced 
new shoot initiation, but had little o r  n o  effect on  plants during fruit production o r  in 
their response to  defoliation. 
3. Shading of plants during growth initiation had n o  detectable impact on  shoot 
growth o r  storage use. Growth initiation caused a reduction in corm mass but n o  
change in per  cent T N C  in corms. Reproduction was associated with reductions in 
both corm mass and per cent TNC.  
4. Artificial defoliation had little impact on  the  utilization of existing carbohydrate 
stores. 
5. I t  is concluded that  carbohydrate storage in corms was relatively unimportant for 
the  recovery of plants from herbivory in comparison to its role in growth initiation 
and reproduction. 
Key-words: Accumulation, artificial defoliation, corms, herbivory, phenology, reproduction, reserves, 
storage, total non-structural carbohydrates 
Functional Ecology (1992) 6 ,  575-581 
Introduction 
Storage organs are a common attribute of perennial 
herbaceous plants, yet few studies have considered 
their ecological function (Chapin, Shultze & Mooney 
1990). Storage structures represent reserves and thus 
directly compete with the allocation of resources to 
other plant functions; consequently the cost of 
storage can be high and ecologically important. If 
storage competes with resource allocation to leaves, 
it will have a direct cost associated with the allocation 
of reserves away from leaves and an indirect cost 
associated with the loss of the compounded benefit 
(i.e. future growth) which accrues to resources 
allocated to leaves (Bloom, Chapin & Mooney 1985). 
Storage does not serve as a reserve function if it does 
not compete with other plant functions (Chapin etal. 
1990). Resources can simply accumulate in storage if 
their external supply exceeds the immediate demand. 
As reserves, storage organs may: (1) support the 
initiation of growth following seasonal dormancy; (2) 
support reproduction; or (3) allow plants to respond 
to tissue losses to herbivores, pathogens, or catas- 
trophe (risk aversion). These reserve functions are 
not mutually exclusive and have the common feature 
that reserves provide resources when demand out- 
strips external supply (Chapin et al. 1990). Most 
evidence for functions 1 and 2 comes from observa- 
tional studies of seasonal variation in nutrient and 
carbohydrate levels in storage organs (e.g. Fonda & 
Bliss 1966; Chapin, Johnson & McKendrick 1980; 
Abrahamson & McCrea 1985; Defoliart et a[. 1988). 
Some manipulative studies have emphasized the 
importance of storage organs for reproduction (Lub- 
bers & Lechowicz 1989; Zimmerman 1990). 
However, other studies have found no effect of 
artificial reductions in storage on plant growth or 
reproduction (Lovett Doust 1980; de Kroon, 
Whigham & Watson 1991) suggesting that storage 
organs are important for risk aversion or that they 
represent accumulation and do not function as 
reserves. 
We have investigated carbohydrate storage in 
corms of the woodland orchid Tipularia discolor 
(Pursh) Nutall with respect to the three proposed 
reserve functions. Approximately 80% of biomass is 
allocated to below-ground storage in T. discolor 
(Whigham 1984). We focused on carbohydrate 
storage because Whigham (1984) found that corms 
were only marginally important for nutrient storage 
576 and that direct uptake and retranslocation of 
J. K. Zimmerrnan nutrients from senescing tissues played the largest 
& D. F. Whigham role in supporting growth initiation and repro- 
duction. 
A role of carbohydrate storage in support of new 
growth in T. discolor is suggested by the observation 
that growth is initiated in early autumn when light 
levels on the forest floor are still very low (Whigham 
& O'Neill 1991). Flowering and fruit production 
occur in the summer when plants are leafless and 
occur at the expense of future corm growth (Snow & 
Whigham 1989). The role of current storage in 
mitigating reproductive costs is unknown. Finally, 
Whigham (1990) showed that artificial defoliation in 
T. discolor has a much greater impact on below- 
ground than above-ground biomass, suggesting that 
below-ground stores were being utilized to maintain 
long-term growth. Thus, there is evidence to suggest 
that carbohydrate storage in the corms of T. discolor 
may play a role in all three proposed reserve func- 
tions. Our objective was to quantify the relative 
importance of stored carbohydrates in each of these 
functions. 
Materials and methods 
Tipularia discolor is a common winter-evergreen 
orchid in woodlands of the south-eastern USA (Luer 
1975). Plants produce a single leaf during Septem- 
ber-October and a single corm is produced at the 
base of each leaf. Leaves senesce in May, flowering 
occurs in early August, and fruits mature by late 
September (Whigham & McWethy 1980). The 
shallowly rooted corms persist for one to several 
years such that plants have a standing crop of two to 
five corms (Whigham 1984; Efrid 1987). We refer to 
corms of differing ages as C1, C2, C3, etc., beginning 
with the youngest, fully formed corm. Vegetative 
reproduction occurs when large flowering plants 
produce a second leaf and corm on the youngest, fully 
formed corm (Snow & Whigham 1989). New plants 
become independent after several years when the 
connection between old and new plants disintegrates 
(Efrid 1987). Herbivores, usually whitetail deer 
(Odocoileus virginianus Zimmermann), frequently 
eat entire leaves of T. discolor and plants do not 
resume growth until the following autumn when a 
new leaf is produced (Whigham 1990). Our studies 
were conducted in two upland hardwood forest sites 
at the Smithsonian Environmental Research Center 
near Annapolis, Maryland, USA (see Whigham & 
McWethy 1980 and Whigham 1984 for details). 
EFFECTS OF SHADING AND CORM SEVERANCE 
ON GROWTH INITIATION AND FRUIT SET 
On 31 August 1989 clones of T. discolor were located 
that contained at least one independent vegetative 
and reproductive individual. Reproductive plants 
had flowered and initiated fruit filling. Pairs of plants 
were assigned, in turn, to four treatments: (1) 
unmanipulated control plants; (2) shading to 6.5% of 
ambient understorey total daily photosynthetic 
photon flux density (PPFD); (3) severance of the 
youngest, fully formed corm (C1) from older corms 
(C2, C3, etc.); and (4) the combination of treatments 
(2) and (3). Each treatment was replicated four times 
at each site. Shading was provided by three layers of 
shade cloth 90cm in diameter suspended 25cm above 
the ground. 
At the start of the experiment each plant was 
gently excavated and the length (L),  width (W), and 
height (H) of each corm was measured with calipers. 
A separate set of harvests indicated that In dry mass 
of C1 = -3.670 + 0.073 x L + 0.121 x H (r2 = 0.47, 
n = 29). The In dry mass of C2 and older corms = 
-5.729 + 0.052 x L + 0.075 x W + 0.185 x H (r2 = 
0.73, n = 92). The severance of older corms 
removed, on average, 29.9% (SD = 14.9, n = 30) of 
the estimated dry corm mass per plant. 
Experimental plants were harvested on 11 and 12 
October 1989. Data were recovered for 27 of the 
potential 32 pairs of plants. Following measurement 
of the length of the primary leaf (petiole + blade), 
plants were separated into new shoots (leaves, 
developing corms and new roots), secondary shoots 
(if present), C1 corms, C2 and older corms (if 
present), old roots, and (if present) reproductive 
structures. These were dried in a forced-air oven 
(SC-350, Grieve Corp., Illinois, USA) at 60?C for at 
least 72h before weighing. Dry masses of primary 
and secondary shoots were combined for data 
analysis. 
Corms and leaves were stored in a desiccator 
before being ground (Wiley Mill, A.H. Thomas Co., 
Pennsylvania, USA) to pass through a 40-mesh 
screen. Ground samples were sealed in plastic vials 
until analysis for total non-structural carbohydrates 
(TNC) using the anthrone method (Spiro 1966) 
following extraction and hydrolysis of starch using 
amylase (Smith 1981). Where sample size permitted 
(>90% of samples), analyses for TNC were run in 
duplicate. Roots contained <lo% TNC and are only 
a small component of plant biomass (Whigham 
1984). 
DETERMINATION OF FLOWERING 
On 17 April 1990 we located and marked 47 indepen- 
dent plants at both sites, measured the size of their 
leaves (length x width) and determined whether 
these plants had developed floral primordia on the C1 
corm. On 7 June 1990 we harvested a random subset 
of plants within a range of leaf sizes which included 
plants with (n = 8) and without (n = 8) floral 
primordia. Among flowering plants, noticeable elon- 
gation of the floral primoridia had begun with the 
577 exception of two plants in which flowering is referred 
Carbohydrate to as 'arrested'. Harvested plants were separated into 
storage in C1, C2 plus older corms, and roots, and dried, 
Tipularia weighed, stored, ground and analysed for TNC as 
previously described. 
EFFECTS OF CORM SEVERANCE ON RECOVERY 
FROM HERBIVORY 
Beginning on 24 October 1989 we excavated and 
measured corms and leaves of 36 plants at each site. 
Plants were randomly assigned to three treatments: 
(1) unmanipulated; (2) artificial defoliation; (3) 
artificial defoliation plus severance of C1 from older 
corms. Artificial defoliation was conducted by 
removing the leaf blade, mimicking the pattern of 
natural herbivory (Whigham 1990). The severed 
leaves were used to estimate the leaf area (LI-3050A 
area meter, Li-Cor, Lincoln, Nebraska, USA) of the 
unmanipulated plants from measurements of blade 
length (L) and width (W). The resultant equation 
was V'leaf area = -0.068 + 0.270 x L + 0.611 x W 
(r2 = 0.98, n = 47). The corm severance procedure 
removed an average of 27.6% (SD = 0.15, n = 18) of 
the total estimated corm mass. 
Plants were harvested nearly 1 year later (10 
October 1990). Plants that had flowered (relatively 
few) and those which suffered natural herbivory were 
not sampled. Eight unmanipulated plants were 
recovered from the experimental design. Ten 
randomly chosen plants from the artificial defoliation 
treatment were harvested for comparison, along with 
eight plants in which defoliation and corm severance 
were combined. Plants were measured for leaf area, 
dry mass of new shoots (leaf + developing corm + 
new roots), C1 corm (produced during the course of 
the experiment), C2 corm (most recent, fully 
developed corm at the start of the experiment), Cg 
and older corms (if present) and roots, and TNC in 
corms. 
STATISTICAL ANALYSES 
Where possible, plant responses to experimental 
treatments were examined using analysis of cova- 
riance (ANCOVA; SAS 1987). In some cases no 
available covariate improved the statistical models or 
the data did not satisfy the assumptions of ANCOVA. 
Where ANCoVA was used, corrected treatment 
means are presented and the covariate used is 
indicated in the text. Biomass data were subject to In 
transformation and per cent TNC data were subject 
to arcsin-square root transformation. Because site 
distinctions were of secondary importance to the 
overall results, site effects were included in ANOVA 
and ANCOVA models but interaction terms involving 
site and other treatments were not. Site effects were 
only occasionally significant. 
Fig. 1. Interaction plots for the responses of vegetative (V) 
and reproductive (R) plants of Tipularia discolor during 
shoot development. Experimental treatments are: control 
(M); shading, no severance of older corms from the 
youngest (C,) corm (U); corm severance, no shading 
(0- - -0); and shading plus corm severance ( e -  - - e ) .  (a) 
Mean leaf length, corrected for In initial total corm mass 
(estimated, see Materials and methods); (b) mean shoot 
mass, which includes secondary shoots (if present), correc- 
ted for In initial total corm mass; (c) mean TNC (% dry 
mass) in C, corm, corrected for In initial mass. 
Results 
EFFECTS OF SHADING A N D  CORM SEVERANCE 
ON GROWTH INITIATION A N D  FRUIT SET 
The responses of plants tc~ the factorial shading and 
corm severance experiment are presented as a series 
of interaction plots (Fig. 1) exhibiting the differences 
in means between vegetative and reproductive plants 
in each treatment. Because vegetative and reproduc- 
tive plants were paired within a clone, the data were 
analysed using repeated measures ANOVA following 
correction for differences in initial plant size. For leaf 
length (Fig. la), the strongest response appeared to 
be due to shading (F1,21 = 104, P << 0.001). Leaves 
of shaded plants were much longer than those of 
unshaded plants owing to aetiolation of the petioles 
(F1.21 = 61, P << 0.001). Leaves of reproductive 
plants were significantly shorter than those of vegeta- 
tive plants (F1,21 = 36, P << 0.001) and plants with 
their older corms removed were shorter than 
unmanipulated plants (F1,2, = 7.6, P = 0.012). In 
addition, there was a significant interaction between 
578 1000 / 
/ J. K. Zirnrnerrnan / 
Initial (estimated) mass of C I (mg) 
Fig. 2. Relationship of initial estimated mass (see Materials 
and methods) of youngest, fully formed corm (C,) to final 
CI mass in vegetative (0) and reproductive ( 0 )  Tipularia 
discolor. For vegetative plants, In final mass = -0.066 + 
1.417 x In initial mass; for reproductive plants, In final mass 
= -1.318 + 0.708 x In initial mass. Dotted line denotes 1:l  
relationship for initial and final CI mass. 
the shading and corm severance treatments (F1,21 = 
5.4, P = 0.030), indicating that corm severance 
influenced the leaf length of shaded plants more than 
unshaded plants (Fig. la). 
Despite the effects of shading on leaf length, there 
was no detectable effect of shading on new shoot 
mass (Fig. lb;  F1,21 = 1.9, P = 0.17). However, shoot 
masses of vegetative plants were significantly greater 
than those for reproductive plants (F1,21 = 106, P 
<< 0.001) and shoot mass was significantly reduced 
by corm severance (F1,21 = 11.1, P = 0.003). There 
was also a significant interaction between reproduc- 
tive status and corm severance (F1.21 = 6.1, P = 
0.022), indicating that shoot masses of reproductive 
plants were more strongly influenced by corm sever- 
ance than vegetative plants (Fig. lb). 
The concentration of TNC in the youngest, fully 
formed corm (C1; Fig. lc), the only corm remaining 
following corm severance, was significantly influ- 
enced only by the comparison of vegetative and 
reproductive plants (F1,22 = 47.4, P << 0.001). 
Similarly, per cent TNC in leaves of vegetative plants 
(20.3% of dry weight) was significantly higher than 
that of reproductive plants (15.2%; F1,19 = 31.8, P 
<<0.001). 
The final mass of the C1 corm (Fig. 2) could not be 
analysed by ANCOVA because the relationships 
between estimated initial corm mass and the final 
mass in vegetative and reproductive plants were not 
statistically parallel (F1,50 = 9.2, P = 0.004). Never- 
theless, the data show a clear separation of vegetative 
and reproductive plants. Reproductive plants exhi- 
bited a stronger decline in corm mass than vegetative 
plants when compared to initial (estimated) values. 
Within groups of vegetative and reproductive plants, 
the effects of shading or severance on C1 mass 
following shoot development were not significant 
(NS). 
DETERMINATION OF FLOWERING 
Plants which had initiated flowering by early summer 
of 1990 displayed a significant negative relationship 
(r2 = 0.94, P < 0.001) between per cent TNC in C1 
and leaf size (Fig. 3a). Values of per cent TNC ranged 
from 64 to 72%. This relationship was NS in vegeta- 
tive plants (r2 = 0.0, P = 0.99) and these plants 
exhibited an average per cent TNC of 66%. The data 
points for vegetative plants fell below and to the left 
of the 99% confidence level CL for the regression 
involving reproductive plants. Data for two plants in 
which flowering was arrested following development 
of a floral meristem fell within the 99% CL of 
reproductive plants. 
The mass of TNC in the C1 corm of plants initiating 
reproduction (Fig. 3b) was statistically constant 
across the range of leaf sizes sampled (r2 = 0.07, P = 
0.57), but there was a significant positive relationship 
for vegetative plants (r2 = 0.60, P = 0.04). Two 
plants which had arrested inflorescence development 
fell below the area of scatter for reproductive plants. 
The single vegetative plant with the largest leaf size 
Leaf size (cm) 
Fig. 3. Relationships of (a) per cent TNC in the C1 corm (in 
radians) and (b) the mass of TNC in C1 with respect to leaf 
size (%'leaf area) in vegetative (0) and reproductive ( 0 )  
Tipularia discolor. In (a) 99% CL are provided for the 
regression for reproductive plants, which excludes two 
plants (0) in which flowering was arrested following 
development of reproductive meristems. The regression 
equation was: per cent TNC (radians) = -0.115 -0.013 x 
leaf size. For (b) the regression equation for vegetative 
plants was: In mass TNC = -8.82 + 1.89 X leaf size. The 
regression equation for reproductive plants was NS and is 
presented only for comparison. 
579 
Carbohydrate 
storage in 
Tipularia 
Fig. 4. Responses of Tipularia discolor to artificial defoli- 
ation (H) and artificial defoliation plus severance of connec- 
tion to old corms (H+S) in comparison to unmanipulated 
plants (C). Plants were harvested 1 year later: (a) Mean leaf 
size (gleaf area), corrected for initial leaf size; (b) mean 
shoot mass (leaf + developingcorm + new roots), corrected 
for In initial total corm mass (estimated, see Materials and 
methods); (c) mean mass of C1 corm, corrected for In initial 
mass of C2 corm; (d) mean TNC (% dry mass) in C,, 
corrected for In initial mass of C2; (e) mean mass of C2; (f) 
mean per cent TNC in C2. Error bars are SE. 
fell well within the relationship for reproductive 
plants (Fig. 3b), but exhibited arelatively low level of 
per cent TNC in the C1 corm (Fig. 3a). 
EFFECT OF CORM SEVERANCE ON RECOVERY 
FROM HERBIVORY 
Responses of vegetative T. discolor to defoliation 
and corm severance (Fig. 4) were analysed using 
one-way ANOVA and two pre-planned contrasts 
comparing control (C) and defoliated plants and, 
among defoliated plants, those in Which older corms 
were severed away (H+S) vs those with unmanipu- 
lated corms (H). Leaf size in the year following the 
manipulations (Fig. 4a) was significantly reduced by 
defoliation (F1,21 = 32.0, P << 0.001), but corm 
severance had no additional influence on the leaf size 
of defoliated plants (F1,21 = 0.13, P = 0.72). 
Similarly, the contrast of new shoot mass (Fig. 4b) of 
control and defoliated plants was highly significant 
(F1,21 = 23.9, P << 0.001), but the difference 
between H and H + S  treatments was NS (FlXzl = 
0.04, P = 0.83). Similar patterns were displayed by 
the mass and carbohydrate storage in corms pro- 
duced during the course of the experiment (C1; Fig. 
4 q d )  and in the mass of corms present at the start of 
the experiment (C2; Fig. 4e). The overall influence of 
the treatments on per cent TNC in C2 (Fig. 4f) was NS 
(F1.21 = 2.37, P = O.ll),  but the pattern of response 
was similar to that seen for other variables. 
Discussion 
Our results suggest that carbohydrate reserves in 
corms of T. discolor serve multiple ecological func- 
tions but differ in their relative importance for these 
functions. Stored carbohydrates appeared of rela- 
tively equal importance for supporting the initiation 
of new growth and reproduction and were relatively 
unimportant for supporting the recovery of plants 
from herbivory. While some of the observed 
responses may involve nutrient storage, Whigham's 
(1984) previous study of nutrient allocation in T. 
discolor suggested that nutrients present in corms are 
not likely to  serve as a reserve function for growth 
initiation and reproduction. Implications of nutrients 
for responses to herbivory are discussed below. 
The importance of corms to growth initiation is 
underscored by the fact that this was the only aspect 
of reserve utilization that was clearly influenced by 
the severance of older corms from the youngest, C1 
corm. During the 6 weeks when new growth was 
initiated, vegetative plants lost, on average, about 
120mg of mass from the youngest, fully formed (CJ 
corm, about 45% of the initial mass. However, 
following shoot initiation, per cent TNC in the 
youngest corms was 61% in vegetative plants (Fig. 
lc),  not much less than the 66% measured following 
dormancy (measured the following summer, Fig. 3a). 
Yet, a net export of carbohydrates must have occur- 
red to  maintain their concentration in the C ,  corm 
during mass loss. 
Curiously, shoot mass following growth initiation 
was not strongly influenced by shading to very low 
levels of light availability. Shaded plants were aetio- 
lated, which is a typical response to  light stress 
(Salisbury & Ross 1978). In the related winter-ever- 
green orchid Aplectrum hyemale (Adams 1970), 
leaves are physiologically adapted to high light levels 
and low temperatures such that photosynthetic rates 
in the autumn are relatively low. If leaves of T. 
discolor are similar, then the contrast of ambient 
understorey light levels and the shade treatment may 
not have had a large impact on photosynthetic carbon 
gain. These photosynthetic characteristics would 
emphasize the importance of carbohydrate reserves 
for supporting growth initiation in the autumn. 
However, T. discolor may also acquire carbon 
through mycorrhizal associations with saprophytic 
fungi as in some orchids which are achlorophyllous 
(e.g. members of the related genus Corallorhiza; 
Dressler 1981). 
Shoots initiated by reproductive plants were much 
smaller than those produced by vegetative plants, 
supporting the results of Snow & Whigham (1989) 
which show that plants suffer a cost of reproduction 
580 affecting both future growth and reproduction. Our 
J.  K. Zimmerrnan results show that carbohydrate storage alleviates 
& D. F. Whigham these reproductive costs to some degree. Duringfruit 
set and shoot initiation the C1 corms of reproductive 
plants lost an average of 67% of their initial mass (c.  
350mg of corm mass), a greater loss than observed in 
vegetative plants. The loss of corm mass during 
reproduction was coupled with a decline in the 
concentration of storage carbohydrates in both corms 
and leaves. 
While plant size is an important determinant of 
flowering in T. discolor, large plants do not always 
flower (Snow & Whigham 1989). Our results indicate 
that the status of carbohydrate storage in the 
youngest corms (Fig. 3) appears to be an added 
determinant of flowering that is independent of plant 
size (Tissue & Nobel 1990). 
Few studies of wild plants have explicitly con- 
sidered the comparative role of storage in growth 
initiation and reproduction. Several studies of tem- 
perate and arctic perennial herbs suggest that below- 
ground reserves are more important for spring 
regrowth than for reproduction (Fonda & Bliss 1966; 
Bradbury & Hofstra 1977; Mark & Chapin 1989). In 
horticulturally grown tulips, translocation studies 
indicate that growth initiation is supported by stored 
carbohydrates but that subsequent growth and 
flowering are supported by current photosynthesis 
(Ho & Rees 1976, 1977). By contrast, Zimmerman 
(1990) reported a pattern opposite to  this in the 
deciduous, tropical orchid Catasetum viridiflavum. 
Artificial reduction of the number of pseudobulbs in 
C. viridiflavum influenced growth in only the smallest 
plants, but caused a large reduction in the number of 
flowers produced by plants of all sizes. 
Our results suggest that carbohydrate reserves of 
T. discolor are relatively unimportant for their 
recovery from herbivory. After 1 year, defoliation 
caused a reduction in the mass and per cent TNC in 
the newest corm (C1) representing a difference of 
about 140mg TNC between control and defoliated 
plants (Fig. 4c,d). The'effect on existing storage (C2) 
represented at most 25 mg of corm TNC (Fig. 4e,f). 
Thus, artificial defoliation has a much larger impact 
on the production of new carbohydrate storage than 
on existing storage. Moreover, severance of the 
connection to older corms in conjunction with defoli- 
ation had no apparent effect on the ability of the plant 
to respond to herbivory. Thus, while carbohydrates 
stored in corms of T. discolor appear to  play some 
role in the response of plants to  herbivory, their role 
appears to  be smaller than first suspected (Whigham 
1990). 
These results are in accord with other studies which 
have shown a reduction in carbohydrate reserves 
following artificial defoliation, but which also fre- 
quently show that large proportions of stored carbo- 
hydrates remain during regrowth (Chapin et al. 
1990). One explanation for this pattern is that 
regrowth is limited by nitrogen and phosphorus to a 
greater extent than by access to  carbon. During the 
growing season, leaves of T. discolor represent about 
20% of plant biomass, but contain up to 40% of total 
nitrogen and phosphorus (Whigham 1984). Thus, 
defoliation must have a relatively large impact on the 
nutrient budget of these plants. 
We have demonstrated that carbohydrates stored 
in the corms of T. discolor provide a reserve function 
and are more important for supporting growth initi- 
ation and reproduction than for recovery from severe 
herbivory. However, the possibility that some of the 
stored carbohydrates are not reserves, but represent 
accumulation that does not compete with allocation 
to growth or  other processes, needs to be considered. 
This is suggested by the fact that, even when statisti- 
cally significant, removal of the older corms had.only 
a small impact on the plants. If some corm carbohy- 
drates represent accumulation, then plants should 
respond to increased nutrient availability (e.g. artifi- 
cial fertilization) by reducing carbohydrate storage 
(Chapin et al. 1980) as the factors limiting growth are 
brought into closer balance. Similarly, elevated 
nutrient availability should strongly influence the 
ability of plants to recover from herbivory if this 
reponse is limited more by nutrients than carbohy- 
drates as our results imply. 
Acknowledgements 
We thank B. Drake, J .  O'Neill, L. Remenapp and D. 
Tissue for their assistance with this project. W.  Arp  
provided useful comments on a previous version of 
the manuscript. Financial support provided by the 
Smithsonian Environmental Sciences Program and a 
Postdoctoral Fellowship from the Smithsonian Insti- 
tution to JKZ. 
References 
Abrahamson, W.G. & McCrea, K.D. (1985) Seasonal 
nutrient dynamics of Solidago altissima (Compositae). 
Bulletin of the Torrey Botanical Club 112, 414-420. 
Adams, J.S. (1970) Adaptations of Aplectrum hymenale to 
the environment: effects of preconditioning temperature 
on net photosynthesis. Bulletin of the Torrey Botanical 
Club 97, 219-224. 
Bloom, A.J., Chapin 111, F.S. & Mooney, H.A. (1985) 
Resource limitation in plants - an economic analogy. 
Annual Review of Ecology and Systematics 16, 363-392. 
Bradbury, I.K. & Hofstra, G. (1977) Assimilate distribu- 
tion patterns and carbohydrate concentration changes in 
organs of Solidago canadensis during an annual develop- 
mental cycle. Canadian Journalof Botany 55,1121-1127. 
Chapin 111, F.S., Johnson, D.A. & McKendrick, J.D. 
(1980) Seasonal movement of nutrients in plants differing 
in growth form in an Alaskan tundra ecosystem: impli- 
cations for herbivory. Journal of Ecology 68, 189-209. 
Chapin 111, F.S., Shultze, E. & Mooney, H.A. (1990) The 
ecology and economics of storage in plants. Annual 
Review of Ecology and Systematics 21,423-447. 
581 Defoliart, L.S., Griffith, M., Chapin 111, F.S. & Jonasson, 
Carbohydrate S. (1988) Seasonal patterns of photosynthesis and 
storage in nutrient storage in Eriophorum vaginatum L.,  an arctic 
sedge. Functional Ecology 2, 185-194. Tipularia Dressler, R.L. (1981) The Orchids, Natural History and 
Classification. Harvard University Press, Cambridge. 
Efrid, R.B. (1987) Corm growth and clonal propagation in 
Tipularia discolor. MS thesis, Clemson University, South 
Carolina. 
Fonda, R.W. & Bliss, L.C. (1966) Annual carbohydrate 
cycle of alpine plants on Mt. Washington, New 
Hampshire. Bulletin of the Torrey Botanical Club 93, 
26s277. 
Ho, L.C. & Rees, A.R. (1976) Re-mobilization and 
redistribution of reserves in the tulip bulb in relation to 
new growth until anthesis. New Phytologist 76, 59-68. 
Ho, L.C. & Rees, A.R.  (1977) The contribution of current 
photosynthesis to growth and development in the tulip 
during flowering. New Phytologist 78, 65-70. 
de Kroon, H., Whigham, D.F. & Watson, M.A. (1991) 
Developmental ecology of mayapple: effects of rhizome 
severing, fertilization, and timing of shoot senescence. 
Functional Ecology 5,36&368. 
Lovett Doust, J .  (1980) Experimental manipulations of 
patterns of resource allocation in the growth cycle and 
reproduction of Smyrnium olusatrum L. Biological Jour- 
nal of the Linnean Society 13, 155-166. 
Lubbers, A.E. & Lechowicz, M.J. (1989) Effects of leaf 
removal on reproduction vs. belowground storage in 
Trillium grandiflorum. Ecology 70, 85-96. 
Luer, C.A. (1975) The Native Orchids of the United States 
and Canada, excluding Florida. New York Botanical 
Garden, New York. 
Mark, A.F. & Chapin 111, F.S. (1989) Seasonal control over 
allocation to reproduction in a tussock-forming and a 
rhizomatous species of Eriophorum in central Alaska. 
Oecologia 78, 27-34. 
Salisbury, F.B. &Ross, C.W. (1978) Plant Physiology, 2nd 
edn. Wadsworth Publishing, Belmont, California. 
SAS Institute Inc. (1987) SASISTAT Guide for Personal 
Computers, Version 6 edn. SAS Institute Inc., Cary, 
North Carolina. 
Smith, D.  (1981) Removing and analyzing total nonstructu- 
ral carbohydrates from plant tissue. Wisconsin Agricultu- 
ral Experiment Station Bulletin 2107, 1-1 1. 
Snow, A.A. & Whigham, D.F. (1989) Costs of flower and 
fruit production in Tipularia discolor (Orchidaceae). 
Ecology 70, 128G1293. 
Spiro, R.G. (1966) Anthrone assay for neutral sugars. 
Methods in Enzymology, vol. 8 (eds. S.P. Colowick & 
N.O. Kaplan), pp. 3-26. Academic Press, New York. 
Tissue, D.T. & Nobel, P.S. (1990) Carbon relations of 
flowering in a semelparous clonal desert perennial. 
Ecology 71, 273-281. 
Whigham, D.F. (1984) Biomass and nutrient allocation of 
Tipularia discolor (Orchidaceae). Oikos 42, 303-313. 
Whigham, D.F. (1990) The effect of experimental defoli- 
ation on the growth and reproduction of a woodland 
orchid, Tipularia discolor. Canadian Journal of Botany 
68, 1812-1816. 
Whigham, D.F. & McWethy, M. (1980) Studies on the 
pollination ecology of Tipularia discolor (Orchidaceae). 
American Journal of Botany 67, 55&555. 
Whigham, D.F. & O'Neill, J .  (1991) The dynamics and 
flowering and fruit production of two eastern North 
American terrestrial orchids, Tipularia discolor and 
Liparis lilifolia. Population Ecology of Terrestrial 
Orchids (eds. T.C.E. Wells & J.H. (Willems), pp. 
89-101. SPB Academic Publishing, The Hague. 
Zimmerman, J.K. (1990) Role of pseudobulbs in growth 
and flowering of Catasetum viridiJIavum (Orchidaceae). 
American Journal of Botany 77, 533-542. 
Received 7 May 1991; revised29 November 1991; accepted5 
December 1991