BIOLOGY OF THE GEOPHYTIC LILY, TRITELEIA LAXA (THEMIDACEAE), IN GRASSLANDS OF THE NORTHERN SACRAMENTO VALLEY

2006·16.7 MB·English

by  Robert A Schlising

Madrono, Vol. 53, No. 4, pp. 321-341, 2006
BIOLOGY OF THE GEOPHYTIC LILY,
TRITELEIA LAXA (THEMIDACEAE), IN GRASSLANDS OF THE
NORTHERN SACRAMENTO VALLEY
Robert A. Schlising and Scott A. Chamberlain'
CA
Department of Biological Sciences, California State University, Chico, 95929-0515
Abstract
Triteleia laxa (Themidaceae), a liliaceous geophyte common in California at lower elevations with
mediterraneanclimate, was studiedduring 1998-2004at foursites in the northern Sacramento Valley.
New individual plants were randomly selected each year at one to four of these grassland or open
savanna sitesto provideacomposite study ofthebiologyand lifehistoryofthe plant. A mature plant
produced two leavesafterfall rainsbegin, a newcorm on topofthe shrinkingoldcorm in spring, and
a scape with an umbel of large bluish flowers in late March or April. By seed production in May,
above-ground parts were dead, and the new corm and seeds persisted through summer. Mean leaf
lengths varied from 22 to 28 cm; no correlation was found with amount ofprecipitation. Mean scape
lengths also varied, from 23 to 32 cm; overall, there was a negative correlation with spring
precipitation. Corms offlowering plants had a mean volume of 1.30 cm\ and occurred at a mean
depth of 7.27 cm; corm volume was correlated with scape height and with number offlowers. Dry
massincreasedlinearlyinthenewcormduringthe springgrowth period. Numberofflowersperplant
varied among sites and the mean varied highly significantly among 5 yr (range 8-12.6). Plants
appeared tobelargelyself-incompatible, butproduceda small numberofselfed seedswhen pollinated
by hand. The pollen:ovule ratio was about 3100:1. Mean fruit set (range 50-74%) and mean seed set
(range 40-58%) were low, and varied significantly among sites and years. Seed mass averaged about
1.7 mg, and seeds showed 100% viability. Germination timewas about4 wk afterwetting, with up to
98% of seeds germinating. The seedling was carried downward from the seed by the elongating
cotyledon, ultimately producing a single foHage leaf, a lateral contractile root that shortened in April
(and in situ doubtless pulls the seedling deeper into the soil), and a corm <4 mm in diameter. Soils
occupied in study sites were loams, with 20 to 39% clay particles. This study on T. laxa gives
a reasonablycompletepictureofthebiologyforacommon geophyte, andwehopethat thiswork will
provide impetus for additional studies ofother widespread or local geophytes.
KeyWords: Triteleialaxa, geophyte, mediterraneanclimate, annualgrassland, corm, contractileroot,
low fruit and seed set.
Geophytes are perennial herbs that survive that of most annual herbaceous species (which
unfavorable periods for growth by dying back to persist as seeds during summer) in California's
underground storage organs such as corms, savannas and grasslands (see especially Chiarello
bulbs, tubers, or rhizomes (Rundel 1996). Dafni 1989).
et al. (1981) have categorized geophytes as either The early, classic study of plant life forms
"synanthous" (leavesand flowers produced in the (Raunkiaer 1934) noted that geophytes were well
same season) or ''hysteranthous" (leaves and represented in areas with mediterranean climate.
flowers appearing in separate seasons). Dafni et In (largely mediterranean) California, monocot-
al. also point out that the underground storage yledonous geophytes constitute about 5% of
organs of these perennial geophytes can be native vascular plant species (Rundel 1996).
"annual" (that is, the old storage organ is Recent studies are still cataloging and making
completely replaced by a new organ each growing numerical comparisons of geophytic species
season) or perennial (with tissue added to the occurring within different regions with this
same organ over the years). In the two-season climate (Poches and Cowling 2004; Poches et al.
mediterranean climate, geophytes commonly dry 2005; Parsons 2000; Parsons and Hopper 2003).
up and become dormant aboveground during the For geophytes of western Australia and some
dry summer and then renew growth from the other regions with mediterranean climate. Pate
storage organ when water becomes available and Dixon (1982) have provided detailed in-
during autumn, winter and spring (Rundel formation on the morphology and anatomy of
1996; Parsons 2000). Thus, these geophytic corms, bulbs and tubers, as well as on growth and
species follow a phenological pattern similar to phenology. There is a paucity of this type of
information for California's geophytes. For
' Current address: Department of Ecology and example, no information on geophytes is in the
Evolutionary Biology, Rice University, Houston, TX book "California Annual Grasslands'" (Huen-
77005 USA, e-mail: [email protected] neke and Mooney 1989).
MADRONO
322 [Vol. 53
Many monocotyledonous geophytes of the natural history in the field is incompletely known
'iily type" (Liliaceae sensu lato, as in Hickman for geophytes in California.
1993; FNAEC 2003) grow in grasslands and To augment the sparse literature on California
savannas of the northern Sacramento Valley of geophytes, Triteleia laxa Benthamwaschosen for
California and include species of Calochortus, study. Triteleia, a mainly Californian genus, is
Chlorogalwu, Fritillaria, Odontostomimi, Zigade- composed of15 species, all ofwhich occurwest of
ims, Alliunu Brodiaea, Dichelostemma, and Trite- the Rocky Mountains (FNAEC 2003). Triteleia
leia. {Triteleia and several others of these genera laxa (commonly called Ithuriel's Spear, Grass
are currently placed in families other than Nut or Walley Basket) is a synanthous geophyte
Liliaceae [e.g., see Pires and Sytsma 2002; occurring in the region with mediterranean
FNAEC 2003].) climate in Western North America, ranging from
Several of the geophytic species that grow in San Bernardino County, California to Currey
California's low elevation grasslands have had County, Oregon (Hoover 1941; Munz and Keck
limited information on features characterizing or 1959; Keator 1993). This widespread species is
m
affecting portions their of life cycles (in these found from sea level to 1500 in elevation
grasslands or elsewhere) published or presented (Keator 1993), less commonly in open forest and
in dissertations and theses. For example, re- chaparral, but more commonly in low elevation
production and corm size in Zigadeimsfremontii savanna and grassland ecosystems.
(Torrey) S. Watson was studied in relation to fire In T. laxa, the perennatingcorm first produces
in southern California (Tyler and Borchert 2002), two long, grasslike leaves during the cool and
and its reproduction and survival have been rainy mediterranean winter season, and then
studied in the northern Sacramento Valley a single scape 10-70 cm tall, bearing an umbel
(Mitchelson 1993). For CalocJiortus luteus Lind- of large bluish trumpet-shaped flowers in the
ley, there are descriptions on floral phenology, spring as the rainy season ends. After fruit and
pollen:ovule ratios and insect flower visitors/ seed production, and formation ofa newcorm on
pollinators (Jokerst 1981), features of the pollen top ofthe old corm, leaves and scape dry up, and
grains (Kannely 2003), and aspects ofpollination aboveground activity ceases during the hot, dry
in relation to evolution in the genus (Dilley et al. summer months. With the fall rains, the corm
2000). In Chlorogalum angustifolium Kellogg, initiates root growth and seed germination also
floral biology was studied by Jernstedt (1982), occurs.
and the same author (Jernstedt 1984) published Triteleia laxa plants are notable for their
considerable detail on seed germination and the morphological variabihty, as in height and in
development of seedlings' contractile roots in C. size and color of flowers (Hoover 1941; Keator
pomeridianum (DC.) Kunth. Jernstedt (1980) also 1993). The time of flowering also varies, with
studied flower development and anthesis in C. more-coastal variants flowering several months
pomeridianum, and Stockhouse and Wells (1978) later than elsewhere (Hoover 1941). Corms of
presented information on pollination in this some variants have been successfully introduced
species. Brodiaea californica Lindley has had into thehorticultural trade, and are often listed in
flower phenology, flower visitors, and seed pro- popular gardening catalogs.
duction documented by Doalson (1999). Triteleia Several environmental conditions of impor-
hyacinthina (Lindley) E. Greene has had seedling tance in the horticultural mass production of T. ;
development described by Smith (1930) and Putz laxa for commerce, e.g., those promoting seed
(1992), and much earlier, Dichelostemma pid- germination and cormel (vegetative corm) pro-
chella (Salisb.) E. Greene and several other duction, have been studied, but only in gardens
geophytes had seedling development described and in laboratory experiments (e.g., Fortanier
(Rimbach 1902). Friti/laria plurijlora Benth. has 1969; Han 1993; Han and Halevy 1993; Han et al.
had low fruit- and seed set described by Witzman 1991). Anatomical development of the ovule,
(1991). There is some basic life cycle information embryo sac, and endosperm has been described
included in the mainly taxonomic papers for (Berg 2003), and polyploidy has been documen-
members of Brodiaea, Dichelostemma and Trite- ted (n = 7, 8, 14, 16, 21, 24) amongwild-collected \
leia by Keator (1987, 1989) and Niehaus (1980), populations (Lenz 1966; Davidson 1975; Keator
and especially in the biosystematic study of 1993). According to Han (1993), this species is |
Brodiaea by Niehaus (1971). Keator (1967) has self-incompatible, requiring pollen transfer be- j
i
amassed considerable information from field and tween plants for seed production.
:
garden studies for six species of Dichelostemma However, there are very few details published
i
(e.g., on phenology, corms, flowering, seed pro- about this plant's biology where it is growing in i
duction, and germination), but most ofthis work the wild. Field studies of this plant are particu-
i
is unpublished. Despite all these studies for larly warranted because it represents one of the
geophytic species that can be found in the area most common geophytes of lowland mediterra- ;
ofthe present study, we and others (Rundel 1996; nean California. Itscorms are almost certainly an
Parsons 2000) believe that life cycle biology and important food source (e.g., Anderson and
i
)
2006] SCHLISING AND CHAMBERLAIN: BIOLOGY OF TRITELEIA LAXA 323
Rowney 1998) that may be used by rodents and In 2003 and 2004, a "Park" study area was
other animals. It also represents a large pollen added, within Bidwell Park, on the east side of
and nectar food-source for insects. Chico (39"46'42.8"N, 121°45'10.8"W) at about
The purpose of our study is to document 99 m in elevation. The Triteleia population here
aspects of the biology and ecology for Triteleia grows on a heavy loam of an ancient alluvial
laxa in the northern Sacramento Valley. We terrace above Chico Creek (Red Bluff gravelly
provide basic information on three topics: loam), with about 25% clay particles (Andrew
1
phenology and growth of leaves, scapes, and Conlin, pers. comm.). Vegetation is savanna, with
corms; 2) features ofsexual reproduction; and 3) a sparse overstory of Qiiercus douglasii. Neither
features ofthe seed and seedling stages in the life burning nor grazing by cattle occurred immedi-
cycle. ately preceding or during the study years.
In 2003 and 2004, a "Road" study area was
also utilized, just to the east of Chico along
Study Areas Humboldt Road (39 44'58.8"N, 121 45'10.8"W)
m
at 131 in elevation. Soils are derived from
Areas with high densities of Triteleia were
Tuscan Formation volcanic mudflows, and are
chosen for study near Chico, California, in the gravelly loams with about 20% clay particles,
northern Sacramento Valley. The main study
classified as Typic Haploxeralfs (Andrew Conlin,
area, utilized 1998 through 2004 and referred
to as ''Vina" below, is located about 16 km north opevresr.stcoormym.o)f.QVueegreetuatsiodonugislassaivia.nnNae,itwhietrh bausrpnairnsge
of Chico, east of Cana Pine Creek Road in
nor grazing by cattle occurred immediately pre-
extreme northern Butte County (39 52'46.6"N,
ceding or during the study years.
(1o2wrn5e8'd34a.n5d"Wm)a,naagtedthbey VtihenaCaPHlfaoirnnsiaPrNeasteurrvee Theclimate at all sites is typical mediterranean,
with cool, wet, winters alternating with hot, dry
Conservancy). Grassland covers thegently rolling summers. Most (86%) precipitation (mean
terrain of this preserve. This region is underlain 55.5 cm) occurs as rain from November through
with cemented materials of volcanic mudflow NOAA
origin, washed down historically from the Tuscan April (Fig. 1; 2005).
Formation ofthe southern Cascade Range to the
east (Broyles 1983, 1987). The soils are usually Methods
m
less than 1 deep on top of this cemented
hardpan (Gowans 1967). Soils where Triteleia is Sampling
most abundant are clay loams, with clay particles
ranging from 33-39% of total soil particles, Transects were established subjectively to cross
according to analysis by A. & L. Western through areas of abundant Triteleia plants. Two
Agricultural Laboratories, Modesto, CA. These permanent 30-m transects were established at
soils will be classified as part ofthe Tuscan Series Vina, one mainly for observations on phenology
in the "Butte Area Soil Survey" being completed and the other for measuring and/or destructive
by the Natural Resources Conservation Service sampling of plant parts. Additional (temporary)
(Andrew Conlin, pers. comm.). These grasslands transects were also set up at each study site for
contain about 275 species ofvascular plants, with observations or for destructive sampling. For
about one third of these being exotic species most aspects of study, plants were selected
m
(Broyles 1987; Mitchelson 1993; Oswald 1997). randomly on cross-transects (up to 6 long) at
This main Vina site, at 52 m in elevation, had not right angles to the main transects every 12 m.
been grazed by cattle since the study began in the Individual plants selected for whole-season ob-
spring of 1998 (Oswald 1997), although cattle servation had wire stakes with aluminum identi-
were present for several weeks in April of 2003. fication labels pushed into the soil 10 cm south of
The area was control-burned in early June 1997, the plant base.
the year before field study began, and again in
1998, 2002, and 2003; in each case, this was after Phenology
the Triteleia growing season was over.
In 1998, 1999, and 2000, a second population Above-ground Shoots and Weather Data. Phe-
of Triteleia was studied in the Preserve, about nological measurements were made for 50-60
2.3 km east ofthe main Vina study area, and east plants along a permanent 30-m transect at Vina
of Hwy 99 in southern-most Tehama County for five years. Most plants selected for study
(39°53'38.6"N, 121 58'34.5"W). This site(referred could not be followed for more than one growing
m
to as "Barn"), at 62 in elevation, has soils season due to the burrowing activities of pocket
similar to those ofthe main Vina site. The region gophers {Thomomys bottae) dislodging the iden-
was control-burned in 1998, after the first year of tification stakes and/or destroying the corms;
study, and cattle grazed here several ofthe winter thus, new individuals werechosen each season. In
months each year ofthe study. some cases gopher activity necessitated choosing
MADRONO
324 [Vol. 53
Totalmontlyprecip.
Precip.departurefromnormal
Meanmonthlytemp.
Temp,departurefromnormal
aFriega.s,1f. roAmpparwoexaitmhaetresdtamteioanninmoOnrltahnldy,tCemApefrraotmurOectaonbderto1ta9l97motnhtrhoulgyhpArperciipli2ta0t0i4on(fNoOrAthAe2V0i0n5a).aMnedaBnarmnonstthuldyy
temperature and precipitation departure is the amount deviating from a 30-yr mean.
severalnewindividuals tomeasureduringa grow- measurements was lower on earlier dates because
ing season. scapes did not start growth at the same time.
Start time for measurement varied during the Mean leaf length became shorter late in the
years of study, but usually began soon after growing season, because tips of the leaves
emerging Triteleia leaves were recognizable at the shriveled and dried and broke off; thus, maximum
soil surface in February. In 1999, start time began leaflengths were compared among the years.
in late March, when leaves were already ap- Analysis of scape lengths was done only on
proaching maximum length. Frequency ofobser- four years (2000, 2002-2004), because data in
vation was weekly in 1998. Leaves and scapes 1999 could not be obtained randomly due to
were measured every four days in 1999. Leaves scapes dying before flowering. Leaf and scape
were measured every two weeks, but scapes length measurements were not made in 1998 or
weekly in 2000. Both leaves and scapes were 200L
measured weekly in 2002, 2003, and 2004. Length
of both leaves was measured to the nearest Corms. In early March 1998, corms were
0.5 cm, from ground surface to leaftip. Emerging excavated from randomly selected Triteleiaplants
scapes were measured from ground surface to the commencing scape growth at Vina. Corm height
tip until the developing inflorescence was appar- and two diameters (widths) were measured with
ent, after which the scape length (=height) was calipers to the nearest 0.5 cm, and corm volume
measured from ground surface to the base ofthe was estimated as the volume of a sphere, based
umbellate inflorescence. on the mean ofthe three measures. In latest April
Phenological measurements on leaf and scape here, and also at the Barn site, additional corms
growth at Vina were not started at the same time were dug at randomly selected points, using pairs
each year, but were carried out until the marked of the tallest and the closest, shortest flowering
plants reached maximum scape height at flower- scapes, in order to determine if scape height,
ingin each offive years. The number ofplants on flower numbers, or soil depth of the corm was
which leafgrowth wasmeasured started out at 50 correlated with corm volume. All measurements
to 60 new individuals each year, but was usually were made with the corm coat still present.
lower 6^4 individuals) by thetime offlowering In the growing season of 2002, random corms
(1
because 1) gophers (and possibly rabbits and were excavated every two weeks at Vina to
mice) herbivorized plants or dislodged/toppled determine volume and mass. This time the corm
identifying stakes, and 2) some individuals were coatwas removed and the two parts ofeachcorm
mis-identified (i.e., Dichelostemmci, Chlorogalum, (the plump, developing, new corm and the
and Calochortus leaves look much like Triteleia shrinking remains of the old mother corm pro-
leaves initially). Also, the number of scape ducing the leaves and scape) were measured
2006] SCHLISING AND CHAMBERLAIN: BIOLOGY OF TRITELEIA LAXA 325
separately. In December of 2004, non-randomly thers from the upper and lower tiers were
selected corms were dug at Vina to check volume compared. Each flower that had pollen counted
and mass, and to determine extent of growth in one anther also had the ovules from the ovary
aftertheearliest rains offall; again, thecorm coat counted under a dissecting scope. Pollen grains X
was removed and old and new parts were 6 (6 anthers per flower) divided by the number of
measured separately. ovules gave the estimate of pollen:ovule ratio.
Ratios were determined for each flower before
Breeding System and Pollination being averaged.
In April 1998, unopened flower buds at Vina
At Vina in 1998, inflorescences were bagged were bagged with strips ofmicroscope lens tissue
with sheets of microscope lens tissue to exclude for 18 or more hours to provide samples of
insect pollen vectors, and later examined for fruit nectar. Nectar withdrawn from bagged flowers
and seed set. In 2000, individual flowers were with 10 mu capillary pipettes had sucrose equiva-
bagged with strips of lens tissue for 48 or more lents (SUE) determined in the field with a Belling-
hours. Bags were gently removed, and wooden ham and Stanley pocket refractometer on 17
toothpicks were used to transfer pollen in three April. A few flowers (despite the bagging) had
types of pollinations within and between inflor- thrips present in the corollas that may have
escences: 1) pollen to stigma in same flower contaminated nectar with pollen; samples with
(autogamous), 2) pollen to stigma in separate suspected contamination were not used to
flowers on same plant (geitonogamous), and calculate mean SUE. The volume of nectar
3) pollen to stigma in separate flowers on varied, and although all samples were from
different plants (xenogamous). Treated flowers separate plants, in 11 of 28 plants more than
were re-bagged and examined for seed set in one flower had to be used to obtain a volume
several weeks. Open-pollinated (marked, but not large enough to register on the refractometer.
manipulated) flowers nearby were used as con-
trols. At Vina in 2004, squares of tulle (a green
mm Reproductive Output: Flowers, Fruits and Seeds
net material with about 1 mesh) were gently
tied around 30 inflorescences to exclude flower To determine variation amongplants in flower,
visitors; 30 nearby plants without tulle were fruit, and seed production, flowers per plant,
marked to serve as controls. All flowers were fruits per plant, and seeds per fruit were counted
exTamoineleudcifdoartefrtuhietspaenridodseoefdsstiingm2a-t3icwkr.eceptivity awtheenachplsainttesinwesreeverianl fyreuairts;. nCuomubnetrs woefreflmowaedres
within individual flowers, a colorimetric assay produced was ascertained by adding number of
was carried out at Park to test for the presence of fruits and empty pedicels remaining—the latter
stigmatic peroxidase enzymes. Four flower stages
(n = 25 for each stage) were designated for the indicating abortion ofeither a flower or a young
fruit. Seeds per fruit were determined for two
assay: a) flower buds nearly open (flowers
dried, indehisced fruits from each randomly-
collected), b) flowers recently opened (1-3 d
selected inflorescence. In several years, a dissect-
old), c) flowers with anthers mostly dehisced
ing microscope was used to also count aborted
and with the style bent towards the middle ofthe
flower {2-A d old), and d) flowers recently closed ovules remaining in the dried fruit.
and most without any sign ofseed formation yet
(4-6 d old). Sampling was done between 2^ pm Seed Weight, Viability, and Germination
(PDT) on 30 April 2003, and samples were placed
in a freezer within twenty minutes of the last A collection of mature seeds from about 200
collection; the assay was carried out seven days inflorescences in the Vina population was made
after collection. Peroxidase activity was detected on 1 June 1997. Seed lots were stored in an
in planta with adoption ofmethods used by Blee indoor laboratory at ambient temperature and in
et al. (2003). an uncooled/unheated garage, where summer and
Pollen/ovule ratios were determined for 35 winter outdoor (=field) temperatures of Chico
plants in the Vina population on 31 March 1998. were approximated. Forty seeds from both
Large flower buds with indehisced anthers were indoor- and garage-stored lots were weighed
collected into botanical FAA. In the lab, one individually, on a Sartorius analytical balance in
anther per flower was stripped of pollen with March 1998.
a needle in a 4 dram vial containing 1 ml FAA or In 2003, collections ofmature seeds, each from
1 ml water. Aftervigorous shaking, an aloquat in 100 or more inflorescences, were made at Vina
a dropper was quickly transferred to a hemocy- (22 May), and at the Park and Road sites (27
tometer and a volume of 0.04 ml solution had May). Seedsfrom each sitewerestoredinagarage
pollen grains counted; this number was multi- in Chico where summer and winter outdoor
plied to represent the number ofgrains dispersed temperatures were approximated, until viability
throughout the entire volume of solution. An- tests and outdoor germination tests were made.
1
MADRONO
326 [Vol. 53
Seeds from each site, stored under outdoor Results
i
temperatures, were subjected to viability tests in
December 2003. For each site, 50 "good" seeds Phenology and Growth ofAboveground Parts
(dark-colored and unwrinkled) were placed in |
a Petri dish, on filter paper wetted with distilled Mean maximum leafand scapelengths differed
water, and kept at ambient temperature in among some years (Table 1). There was not an !
a cupboard for 45 hr. The viability of each seed association between precipitation and leaflength, 'j
was then tested with tetrazolium-chloride adjust- based on precipitation ofthe entire winter-spring
ed to pH 6.5 (Baskin and Baskin 1998). growing season (1 October to date of maximum
To evaluate the proportion of seeds germinat- length) or on precipitation occurring only in the
ing, when they germinated, and when the shoot days of the (warmer) study period starting in
reached the soil surface, seeds collected at Vina February (Figs. 2 and 3). There was, however,
and stored under outdoor conditions in 2003 a highly significant (P < 0.0001) negative
were planted in a garden plot in Chico with correlation (r^ = —0.408, n = 153) ofmaximum
prevailingoutdoortemperatures and only natural scape length with precipitation of the study
precipitation. On 30 November, in a raised period (1 February to date ofmaximum length), i
garden box with sandy loam soil, seeds were Raw data on lengths from four comparable '
mm
planted 8-10 deep, in 10 rows of 50 seeds sample dates in the logistic phase ofscape growth
spaced 20 cm apart, on strips ofmoistened, non- in 2000 and 2002-2004 indicated that there was
inked newsprint paper 3 cm wide (so that re- a significant affect ofyear on scape growth rate.
covery ofgermlings would be possible). The first The highest growth rate, in 2002 (10.4 cm per
row of germinating seeds was recovered in four week) was highly significantly greater (P <
weeks (28 December), and additional rows 0.001) than in all other years (Table 1). The
were exhumed every two weeks through 9 April lowest growth rate was 5.5 cm per week in 2004.
2004. In these fouryears, flowering occurred a fewdays
after maximum scape length was reached.
Climatic Data Leaf lengths, but not scape lengths, for 1999
are included in Table 1, because the scapes died '
Monthly climatic data for Orland, CA were just before flowering. By 7 April, effects of an
used to approximate weather at the Vina study unknown pathogen were first noticed as a down-
site, which is about 18.5 km northeast ofOrland ward bending of the scape several cm from the
(NOAA 2005; Fig. 1). Daily measurements of top. In a few days, the inflorescence and portion
precipitation, and minimum and maximum air of the scape above the bend died, leaving the
temperatures at Orland and Chico were used with unopened, or barely-opened flower cluster dan- |
our phenological measurements and seed germi- gling (Fig. 4a). By 2 May, 33 (94.3%) of the |
I
nation experiment, respectively (UCD 2005). original 35 marked plants producing scapes at
Vina had dead non-fruiting scapes due to this j
disease. A similar effect was observed in the Barn \
Data Analysis
population, with dead scapes on 26 (76.5%) of j
Means for many traits measured (e.g., leaf the original 34 plants producing scapes. Phenol-
length, scape length, flowers per plant) did not ogy was not studied in 2001, but the same scape
always have homogeneous variances, so non- death phenomenon was observed in both Vina
parametric (Kruskal-Wallis, Dunnett T3, Tam- and Barn populations.
hane or Games-Howell) analyses were used to
test for differences in these cases. However, since Corms
we usually had large sample sizes, and the j
probabilities obtained from non-parametric tests Conn volume comparisons, and corm depths in
were nearly identical to the P values obtained the soil. Corm volume (mean ± SE) in cm\ when |
from ANOVA and Tukey HSD comparisons, we first measured on 2 March 1998, at Vina, was
have usually reported the latter. Where its 1.62 ± 0.131 (n = 30); allcormscamefromplants
assumptions could be adequately satisfied, AN- that had flowered the previous year. Paired corm
OVA followed by Tukey HSD comparisons was volumes (longest and shortest flowering scapes)
used to test differences. Repeated measures from Vina longest plants (1.76 ± 0.137; n = 30)
ANOVA was used to determine if scape growth and shortest plants (0.91 ± 0.059; n = 30) were
rate during the logistic growth phase varied highly significantly different (P < 0.0001) as were
I
among years. Spearman's rho rank test was used the paired corm volumes for Barn longest plants
wtoerechepcrkeffoorrmecdorruesliantgionSsP.SSMossotftdwaartea (aRnealleyassees 0(.10.8506;±n0.=1 1340;).nH=ow3e0v)earn,dwshheorntelsotnpgl-apnltasnt(0c.o9r6m±s \;i
11.0.1, Chicago, IL), but some analyses were from Vina and from Barn were compared, they
done using JMP IN 3.2.1 for Macintosh (SAS did not differ significantly in volume, nor did the i;
Institute, Inc., Cary, NC). short-plant corms from the two sites (P = 0.589
j
^
2006] SCHLISING AND CHAMBERLAIN: BIOLOGY OF TRITELEIA LAXA 327
in both cases). All ofthese corms combined from
both sites at the end ofthe growing season in late
C q d ^oo >inn A0p.r0i6l1,csmh'ow(end=an12o0v)e.rall mean volume of 1.30 ±
U 03 Depths ofthese Triteleia corms in the soil were
OO
not significantly different in the paired longest
and shortest plants dug in Vina and Barn
populations (long-plant and short-plant corms
combined at a site to compare Vina vs. Barn, P =
Cr^Trn^ ^1^ 00 ^ Of^NJ 0c.o1m1p2a;reVinalal laonndg Bvsa.rnallcosrhmosrt ccoomrbmis,nedP t=o
0.090). Mean depth (±SE) for these combined
corms (from soil surface to top ofthe corm) was
7.27 ± 0.29 cm (n = 120).
Corm volume correlated with scape length and
number offlowers. The correlation coefficient for
^^ rOo i^n or^l corm volume and scape length was 0.612, and for
corm volume and numbers of flowers was 0.610
(both correlations significant at P = 0.01). In
addition, at Vina and Barn in 1998, we found
number of flowers per plant to be strongly
=
correlated with scape length (Spearman's r^
. X) O — 00 o o 0.814, P = 0.001, n = 120).
(N 0^0' rmj- rmt Or4N Corm volume and changes in mass through
the spring. Only the 28 February volume was
significantly different from each of the four
<
volumes measured later (P 0.017), and none
of the March and April volumes differed
a a a a a
< < < < < significantly from each other (Fig. 5). There
was close to a linear increase in volume through
this 8-week study period, and the deviations of
the sample means did not differ significantly (P <
0.001).
The preceding volume and depth data reported
o cd k3 were obtained with the corm coat present on the
outside. However, removing a T. laxa corm coat
showed that the corm present at the beginning of
the warm growing season (i.e., starting in
February) had two parts. The first (referred to
as the "old" corm), shrinks as leaves and scape
lengthen on the growing plant; this older part
in dON —m \do m— was gradually replaced by an enlarging "new"
(^i corm on top of the old corm (not necessarily at
^+1 +1 ^+1 +1 O+1 the same rate in all plants of a population)
00 ON
(Fig. 4b). Dry mass for five dates measured
starting in February 2002 (during the enlarging
of the new corm, while leaves and scapes were
growing), increased from about 210 to 547 mg
(Fig. 5). Masses of the old corm, (not given in
^d— -—e "iSn1d Od-N mFiega.n5s),rawnegriengqufirtoemsaibmioluatr 2t5o teoac4h1 mogth.erA,nwiotlhd
^+1 +1 +1 +1 ^+1 corm is shown below the new corm in Fig. 4c.
\o \o
Newgrowth ofcorms in the wetseason. In 2004,
mm
there was rain early in the wet season (67
from 17 to 26 October, when temperatures ofair
mm
and soil were still warm, and 107 total by
OONn OoO OrOi OO OO NDeocveemmbbeerr20300)4., vFoolrumceor(mmseandu±g SaEt)Vwiansa1.o2n3 ±1
O—N n ri ri 0.08 cm\ All these plants sampled in December
'
MADRONO
328 [Vol. 53
-5 I > -5
1-Feb 15-Feb 1-Mar 15-Mar 29-Mar 12-Apr
Fig. 2. Vegetative phenology for Triteleia laxa shown by mean (±SE) leafand scape lengths at Vina duringfour
years of the study. Daily precipitation and daily mean maximum and minimum temperatures at Orland, Glenn
County (U.C. Davis 2005), are shown below each year's phenological data. Total wet season precipitation
preceding 1 February is indicated in upper left corner for each year. N's are shown adjacent to the point ofeach
sampling date.
>
r
SCHLISING AND CHAMBERLAIN: BIOLOGY OF TRITELEIA LAXA 329
35
^ 2003
30
E 44 —44 25
3. 25 - — f
20 Leaf 44 44 / 25 44
co 44 44 22
15
c
(0 10
15
5 Scape 2
6
0 1 1
40 40
Totalprecip. 1-Oct.2002 Precip. (mm)
35 to 1-Feb.2003340mm Maxtemp. 35
Mintemp.
30 30
O
o 25 h 25
0)
43— 20 \ 20
(0
\ 15 15
E
10 \ 10
5 5
0 0
-5 -5
1 Feb 15-Feb 1-Mar 15-Mar 29-Mar 12-Apr
35
2004
30
25 44 26
20
/ 39
15
Leaf 44 27
10
44
5 Scape ^^^'^"""""^
0 1 1 1 1
40
Totalprecip. 1-Oct.2003
35
to1-Feb.2004290mm
30
O
o 25
0)
i3- 20
->
(0
\ 15
E 10
5
0
-5
1-Feb 15-Feb 29-Feb 14-Mar 28-Mar 11-Apr
Continued.
(n = 30) had a new, green shoot extending and turgid new corms was 150 ± 3.8 mg, and for
upward from the corm (mean length = 6.6 ± the shrunken old corm still attached (i.e.,
0.45 cm), but no shoots had reached the soil functional during early spring plant growth in
surface. Mean (±SE) dry mass for the 30 plump 2004), 31 ± 3.3 mg. Seven of the 30 corms had
MADRONO
330 [Vol. 53
35
1999
30
35 11
25
33 A 1 _4 ^5 16 i
20 Leaf
c
15
c
10 >^19 19 16
13
5 Scape ,^^5 19
0 1 1 1 1 .1
40
Totalprecip. 1-Oct. 1998 Precip. (mm)
35 to 1-Feb. 1999220mm Max temp.
Mintemp.
30
O
o 25
20
I 15
E
10
5
0
-5
1-Feb 15-F^b 1-Mar 15-Mar 29-Mar 12-Apr
Fig. 3. VegetativephenologyforTriteleialaxashownbymean(±SE)leafandscapelengthsatVinain 1999(ayear
scapesdied beforeflowering). Dailyprecipitationanddailymean maximumandminimum temperaturesatOrland,
GFelbernunaCroyuinstiynd(iUc.aCt.edDainviuspp20e0r5)le,ftarceorsnheor.wnNbevlalouwesthaerpehsenhoolwongiacdajladcaetnat.tToottahlewpeotinsteaosfoenapcrhecsiapmiptlatiinogndparteec.eding 1
remains of the old corm from two years earlier other in seed number (P < 0.119). Both produced
(the corm functioning in 2003) still attached highly significantly more seeds than the self-
below the old corm. pollinations (P < 0.0001). In 2004, among the 36
None ofthe corms dug during any ofthe years covered plants, 23 made one or more fruits.
ofthis study possessed additional small cormlets, However, 26 of the 40 fruits total sampled from
!
derived vegetatively from the main corm, that these plants had no seeds, giving a mean (±SE) of
have been reported elsewhere for T. laxa and 1.1 ± 0.3 seeds set per fruit; the 14 plants with
have been used in horticultural propagation; fruits that contained one to six seeds had a mean
however, geographical variation in cormlet pro- of 3.0 ± 0.4 seeds per fruit. The 39 uncovered |'
duction has been noted by (Hoover 1941). plants produced a mean of 16 (±1; range 1—40)
seeds per fruit (n = 78), and all of these plants j
Breeding System and Pollination had at least two fruits with three or more seeds.
Seeds per fruit were very significantly lower in
Compatibility. In 1998, a single flower was covered (autogamous) than uncovered (open
bagged on each of 37 plants; only 1 flower pollinated) flowers (P < 0.0001).
produced a fruit, whereas all plants produced
fruits from unbagged flowers. In 2000, a few Stigmatic receptivity and changes in floral
seeds were produced in both types of hand self- morphology. Flowers remain open for up to four :
pollinations, within flower (autogamous) and or five days. Anthesis occurs from the period '
within plant (geitonogamous) (Fig. 6). Seed when flowers are about to open to about one to
numbers did not differ statistically in these two two days old (flower phase A to C). Stigmas
types of selfings (P = 0.946); the few seeds become receptive following anthesis; thus flowers
observed suggests a high degree of self-incom- are protandrous. Anthers are located towards the
|
patibility in this population. As indicated in center of the flower during anthesis, after which J
Fig. 6, hand cross-pollinations (xenogamous) and point the stamens reflex towards the perianth, I
open-pollinated controls did not differ from each separating the anthers from the stigma. The
i