Tillandsia usneoides (L.) L.
Literature references:
*move your mouse pointer over the page numbers to see comment
Comments:
- THE ECOLOGY OF SPANISH MOSS (TILLANDSIA USNEOIDES) : ITS GROWTH AND DISTRIBUTION by R. E. GARTH , Northwestern State College, Natchitoches, Louisiana in Ecology 45(3): 470-481. 1964
Abstract. The growth pattern of Spanish moss is one of alternately dominant dichotomous forking (scorpioid dichotomy). The non-dominant branch at each fork is a leaflike branch. The flowers, which are terminal on the pendent plants, are produced in South Georgia from the middle of April to the first of June. The subsequently formed capsule remains closed for 6 months and its remnants remain in the same position for a year. Spanish moss has a positive geotropic response when oriented in a horizontal position. The formation of several vertical plants on old horizontal stems is a means of vegetative reproduction. An almost direct relationship is evident between terminal growth and the percentage of solar radiation. Atmospheric moisture alone will not support growth; experimental plants die in 3-4 months with natural humidity, but no rain.. Rate of growth could not be correlated with moisture or temperature, but it was adversely affected by shade. Although Spanish moss does not appear to favor any one host, it is not often found on pines. Possible explanations for this scarcity are that pines are self-pruning, the dense leaves are a formidable barrier to wind-blown entry, and the proportion of intercepted rainfall is greater in pines than in broad-leaved species.
The distribution of Spanish moss in the United States, limited to the Coastal Plain of the southeastern states, ranging from Texas to Virginia, may be related to major storm paths which arise in Mexico or cross other storm paths which arise there and move laterally over the Coastal Plain. Some of the plants may have been carried into Florida by storms from the islands in the Caribbean Sea. Spanish moss is able to survive and produce viable seeds outside the Coastal Plain, and ability of the species to survive in habitats completely dissimilar to those of the Coastal Plain areas is shown by its distribution in various physiographic regions of South America. The close relationship between Spanish moss and ponds in Georgia appears to be the result of unfavorable conditions elsewhere (such as fire and lumbering which remove suitable substrates), rather than conditions provided by the ponds.
INTRODUCTION
Spanish moss (Tillandsia usneoides L.), reported to have arisen in the Peruvian Andes (Smith 1934), is a well-known plant of the southeastern United States (Fig. 1), with a range
extending through Central America south to Argentina and Chile (Smith 1939). Spanish moss appears to be relatively independent of edaphic factors, relying upon atmospheric conditions, directly or indirectly; for moisture and mineral nutrients. The plant has no absorbing root system, but has a layer of absorbing scales over the entire plant, similar to those found on other mem bers of the Bromeliaceae.
Schimper (1885) and Mez (1904) studied in detail both the anatomy and physiology of water absorption of this plant. Schimper indicated that the outer disk of the absorbing peltate scales was merely a covering which opened when wet and closed when dry. Mez, however, stated that the outer walls of these openings acted as tiny pumps drawing in any water which was on the outer surface. Steinbruick (1905) disagreed with Mez and believed the action of the scales to be entirely dependent upon the cohesive power of water. Aso (1910) investigated the mechanism of absorption of salts, and basing his conclusions on the action of LiNO 3, concluded that salts could pass into the plant through the scales.
Billings (1904) has provided one of the most complete reports of the ecology, anatomy, and reproduction of Spanish moss. His study was confined largely to field observations but included a few experiments on relative humidity and seed germination. He indicated the role of atmospheric humidity and stem run-off of the supporting tree as water sources and stated that the plant seemed to have a "preference" for sunny places. Pittendrigh (1948) also pointed out the strong light demand of bromeliads. He found these plants received their moisture from rain and dew but were not affected by humidity. Penfound and Deiler (1947) stated that Spanish moss "prefers" a well-lighted but moist habitat and that the plants are confined to the "water-side" of trees on stream banks in some areas. They determined the ability of Spanish moss to absorb moisture under various conditions of relative humidity and demonstrated that it could withstand extreme drying, retaining a high percentage of moisture when placed with a strong desiccant.
Spanish moss appears to grow better on dead trees than lives ones . (Wherry and Buchannan 1926) and seems to have a decided "preference" for trees growing in calcareous soils (Harper 1906). Mez (1904) believed that the plant is limited to drier habitats due to a need for regular periods of desiccation. The genus Tillandsia has been called the most primitive and the most xerophytic of the Bromeliaceae (Pittendrigh 1949).
Schimper (1888) found the first seedlings of Spanish moss in Venezuela. Many seedlings have been found in Louisiana (Billings 1904), but attempts to grow them from seeds have been unsuccessful (Billings 1904, Penfound and Deiler 1947). The latter investigators suggested the possibility of the need for an after-ripening period for germination. The only successful germination attempt was reported by R. A. Lewin (private communication, 1953) who grew seedlings using an algae culture medium. He reported that seedlings developed roots and that a 2% sucrose solution tended to aid in root formation.
Thorold (1952) has used the number of vascular epiphytes present as a climatic indicator of an area in relation to disease-control problems, and the altitude of growth of bromeliads in trees has been used to indicate the general atmospheric conditions in a jungle (Foster 1945). Hosokawa (1950) has proposed an epiphytic-quotient using the number of epiphytes present in a given area as an indication of the climate. These usages of epiphytes assume that plant presence is an indication of suitability of the habitat and do not take into account the differences between optimal and suboptimal conditions.
The present study investigated (1) the rate and pattern of growth of Spanish moss; (2) the effect of selected environmental factors on growth; (3) the present distribution in the United States; (4) the distribution patterns. Observations also were made on the life cycle, particularly with reference to sexual reproduction. Field work was begun in January 1953 and continued through May 1954.
The experimental work was done at Emory University and in the vicinity of the Emory University Field Station, Newton, Georgia. Field trips were made to other parts of Georgia, South Carolina, North Carolina, Florida, Alabama, Mississippi, and Louisiana. Distribution data were accumulated from herbarium sheets, published records, personal observations, collections, and correspondence. The nomenclature for trees followed that of Small (1933).
The writer wishes to express his gratitude to Dr. Robert B. Platt for his help and suggestions during his direction of this work, which is based on a thesis submitted to Emory University in partial fulfillment of the requirements for the doctorate degree.
GROWTH AND REPRODUCTION Growth pattern
Spanish moss has a scorpioid dichotomous growth pattern in which stem development alternates with successive nodes, the left branch being dominant at one node, the right branch at the next (Fig. 3). The alternate fork at each node, opposite the dominant branch, is a leaf-like branch. Secondary leaf-like structures appear later at these nodes in the center of the fork. The first of these secondary structures is attached to the dominant branch of the fork and is carried with it, -as the dominant branch elongates. This structure then becomes one side of a new fork and in turn begins to elongate as the dominant branch. At the same time new short branches appear at the preceding forks where they remain, or as will be pointed out later, may become a part of a side branch.
With the exception of the last complete node and internode, there is an internal bundle of sclerenchymatous fibers throughout the length of the stem which gives the stem strength and contains the poorly developed -vascular system. As the plant elongates distally, it dies proximally. The parenchymatous cortex and leaves drop off, leaving only the central core of fibers as an attachment to the substrate. The portion of the plant alive at any one time is usually less than 45 cm long, frequently between 15 and 20 cm in length. The apparent great length of the plants in the festoons is due to the numerous shorter plants which overlap each other. The three or four terminal internodes continue elongating concurrently until each reaches its maximum length and this occurs in a sequence, beginning with the oldest. The plant is continually adding new internodes terminally so that there are always several elongating internodes during the growing season.
In addition to the growth described above, the terminal internodes often exhibit a loose spiraling. The internode branching to the left spirals to the left in a counterclockwise manner and the opposite internodes spiral to the right in a clockwise manner. This spiraling is eventually lost as far as these particular internodes are concerned, possibly due to the weight of the new growth hanging below.
Specimens from many widely separated areas in this country and from Central and South America showed no appreciable differences in growth habit consistent with any distributional
pattern. There were some variations in size, both in diameter and length.
Various combinations of the number and length of leaves were tested to find some easily measurable growth index of Spanish moss. No satisfactory combination was found. The degree of variation within a festoon was as great as that found between habitats. The average amount of terminal elongation for a group of specimens, however, appeared to be a reliable measure of the response of the plant to its habitat. The point of reference for this measurement is the terminal flower or the subsequently formed capsule.
Occasionally, as a result of wind, the distal tip of a plant will become entangled with some object such as another plant, a tree limb, or a piece of wire. This entangling is aided by the spiraling of the plant. As a result, the plant and often the entire festoon is in a horizontal position rather than the usual vertical one. Often only one or two of the plants of a festoon become entangled, leaving the terminal ends of the other plants pendent. The entangled plants take on a new pattern of growth. The nodes retain their meristematic properties and one of the short branches at each node begins to elongate, carrying with it the adjacent leaf-life branch. From this point the side branch continues to elongate in the same manner as the terminal shoot in a pendent plant. These side branches are capable of flowering in season after they are three internodes long. Several of these lateral shoots may appear at the same time and continue to grow simultaneously at approximately the same rate. In effect, each of these side branches is a new plant and the old connecting stem dies, leaving only the central core of fibers to connect the new plants (Fig. 4).
To investigate this apparent geotropic response, four cypress pin-wheels were constructed in such a way that the eight radiating arms, each 38 cm long, were 45º apart (Fig. 5). They were mounted on upright posts l 1/2 m high so that two of the arms were in a vertical position. A single plant was laced to each of the arms, with the growing tip distal to the center of the pinwheel. All plants were as nearly similar as possible, lacking any signs of lateral growth other than one leaf-like branch and one short branch at each node. They had all been growing in the same location in a vertical position. On June 15 the pin-wheels were placed in the partial shade of a grove of oaks (Quercus virginiana Mill.). The number of lateral growths (leaf-like branches and branches) at each node and the number of internodes on the lateral shoots were compared after 15 weeks.
In the upper (180º) and lower (0º) vertical arms, 54% of the internodes had developed structures in excess of the typical two (Table I ) . The 45°, 90°, and 135° arms, however, had 57%, 77%, and 59%, respectively, of the nodes with leaf-like branches in excess of the normal pair. All five of the angles showed some production of new internodes at these points. None of the vertical plants produced more than one internode, whereas at least two were produced on each of the other angles in several plants.
Sexual reproduction
About March 1 in Baker County terminal growth slows appreciably and flower buds appear (Fig. 6.A), replacing the secondary branches which would normally appear. As the stem elongates the dominant branch of the terminal fork carries the flower bud with it and also the new leaf-like branch which is formed on the inner side of the bud (Fig. 6-C). By the time the flower opens it is terminal on the plant and flanked by the two leaf-like branches (Fig. 6-D). One is the tip of the dominant branch and the other is the new leaf-like branch which was carried along with the flower bud. After flowering, a new branch is formed at the next to the last node, and it elongates as if it were the dominant branch of the terminal fork. The new dominant branch carries a young leaf-like branch with it, and the normal growth pattern is resumed. As a result, the flower is bypassed and the once terminal flower is now in a pseudo-axillary position. There are now two consecutive dominant branches on the side opposite the flower.
The flowers of Spanish moss first appear near the middle of April in southern Georgia and persist until early June. There is only one flower on a plant, except in a few cases when "doubles" appear, and every plant does not flower every year. Each individual flower lasts about 4 days. In 1953 the first flowers at the Emory University Field Station were noted on April 24; at the same time plants in central Florida were also flowering. One week previously, plants examined in North Carolina and South Carolina within a 180-km (110 miles) radius of Wilmington, North Carolina, were not flowering. Plants in the greenhouse, however, at the University of Tennessee and the University of North Carolina, had been flowering for 2 weeks prior to this time. In 1953 the flowering period appeared to end in southwest Georgia on June 8, and no flowers were observed for 18 days. On June 26, however, three more flowers were found.
Of 30 flower buds dissected in the laboratory, all but four had pollen grains on the stigma. Five of these had pollen tubes penetrating the style. All flowers which were allowed to open in nature and were subsequently examined had germinating pollen grains on the stigma.
Eight flower buds were enclosed within individual cellophane bags and four were examined after the flower had wilted. None showed developing pollen tubes, but all had self pollen on the stigma. Of the four which were not touched for 3 months, one developed an aborted capsule and the others rotted at an early stage. The average size of ten pollen grains, one from each of ten plants, was 28.0 by 24.5 µ. No pollination agent was noted despite continued observation.
Capsule opening
The three-locular, septicidal capsule becomes evident near the end of June but does not open until the following December or January (Fig. 7). In January 1954 100 capsules, chosen at random, were examined. Eighty per cent of the capsules contained between 11 and 20 seeds each, with a mean of 13. The range of seeds per capsule was from 2 to 23, and the median was 16.5.
Seedlings
The seedlings (Fig. 8) are small green Yshaped plants with non-absorbing root-like holdfasts. The seeds usually are blown away from the capsule soon after it opens but may remain in the opened capsule and germinate there. Seedlings were seen from the first week in January in 1953, but in 1954 were not seen until March 1. Seedlings are observed soon after the capsules open, indicating little if any dormancy.
Seedlings, as well as seeds, may be carried by the wind to new locations by means of a mass of hairs which function like a parachute. These hairs are covered with tiny barbs which aid the seeds and seedlings in attaching to each other, to mature . plants, festoons, bark, etc. Fifteen seedlings were , counted growing on the bark of a single pine (Pinus palustris Mill.) (Fig. 9), and 28 seedlings were counted on the trunk of a live oak, all within 2 m of the ground. In February and March hundreds of seedlings may be found in a single tree growing on festoons of Spanish moss. One vigorous seedling was found attached to a tree trunk which was approximately 250 m away from the nearest Spanish moss.
THE ENVIRONMENT
Moisture requirements
To investigate sources of moisture 15 cubical wire baskets (Fig. 10), made of 21-gauge hardware cloth and coated with asphalt, were filled loosely with Spanish moss. The baskets, 15 cm on a side and open to the top only, were suspended 1 1/2 m above the ground. This condition was abnormal with respect to crowding, light, and orientation of the plant, but the comparison was believed valid since all baskets were similar with respect to these factors. The plants were all taken from the same tree and no appreciable amount of terminal elongation had taken place since the time of flowering. Twelve baskets were shaded from above by an umbrella made of plastic screen. Three of these baskets were placed over water in a swamp, the remainder being placed under two live oaks which partially shaded them. Of this latter group, three were watered ten times per week, three watered five times per week, and the remaining nine, including the uncovered ones and those in the swamp, were not watered. Watering was accomplished by splashing with well water from an Erlenmeyer flask until the plants appeared saturated. All plants had access to atmospheric moisture, and the uncovered ones received the benefit of rain and dew. Dew was not observed on any of the covered plants.
The experiment lasted 16 1/2 weeks, from June 11 to October 5, 1953. Response to the various treatments was measured by the, amount of terminal elongation, the general appearances of the plants and the number of apparently viable capsules formed. Sixty specimens were used for terminal elongation measurements from each series, 20 from each basket. The coefficient of abmodality (significance of the difference of the mean) was calculated (Table II) by formulae taken from Simpson and Roe (1939, pages 192 and 195), and the probability of a significant difference was determined.
Series A, which was watered ten times per week, was significantly different from the other series. The remainder were statistically inseparable. The rainfall was not heavy at any one time but scattered over the entire period.
There were certain subjective differences between the series. The plants in series A were healthy and vigorous and the capsules were nor mal. These plants were indistinguishable from those growing in the surrounding tree. The plants in series B, watered five times per week, were also healthy but had not grown as much as those in the nearby trees. The capsules appeared normal. The plants in series C over the swamp were not watered and while they showed a little growth, did not appear as vigorous as those in series A and B. They were more brittle, and the capsules were shorter and abnormal in appearance and a few were moldy. Series C plants appeared to benefit from the open water when they were compared to series D which were over land and were likewise not watered. Series D plants were very brittle. The majority of these capsules did not mature and those which did were stunted compared to those in series A and B. The plants were abnormally grey with little evidence of the green color peculiar to healthy growth in Spanish moss. The uncovered plants in series E were healthy and the new growth was green. The capsules were long and appeared healthy.
Twelve festoons of Spanish moss collected throughout Florida were placed on a screened porch, open to the east, on April 27, 1953, and were still alive on July 6, 1953, although no measurable terminal growth had taken place. On August 14 of that year, all were dead. Plants were assumed to be dead when chlorophyll was no longer evident. During this period, their sole source of moisture was that which they were able to obtain from the atmosphere. The relative humidity on the porch was between 40% and 80% most of the time.
The planks which were placed over the swamp, but received no rain, did not grow well. They did, however, show some differences which indicated that they were under better conditions than comparable plants not over the swamp. The plants placed over the swamp grew as well as those receiving only rainfall. It would seem, therefore, that frequent rains are required to provide optimal conditions for the plant, but in the absence of these frequent rains, high relative humidity in combination with fairly frequent rains is satisfactory for optimal growth. Penfound and Deiler (1947) have indicated that Spanish moss is capable of utilizing atmospheric moisture when the relative humidity is high, but their experiments do not indicate whether the amount of moisture so acquired is sufficient to sustain the life of the plant indefinitely.
Solar radiation
Shade racks covered with different amounts of cheese cloth were used in a limited experiment to test the influence of solar radiation on growth. The growth of the plants was almost directly proportional to the measured visible solar radiation (Fig. - 11). Terminal elongation of the plants grown at less than 16 1/2 % of full sunlight was very poor.
Field measurements of temperature and relative humidity effects
Mossy Pond in Baker County, Georgia, was :chosen for an experimental transect along which to study Spanish moss distribution and abundance and the effects of moisture, light, and temperature. It is a solution pond with a relatively impervious bottom and has 15.2 permanent water acres, 26.2 acres of tributary area, and reaches 4.5 ft maximum depth at the overflow stage (Hendricks and Goodwin 1952). Prominent trees around the pond include Taxodium disticum (L.) L. G. Rich., Nyssa biflora Walt., Diospyros virginiana L., Pinus palustris Mill., and Quercus virginiana Mill.
Seven stations, located outward from the periphery of the pond, represented a cross section of the extreme and average conditions under which the plant was growing. Each station was 4 m in diameter, with an instrument board in the center On this board were mounted a maximum-minimum thermometer and two Livingston spherical atmometer bulbs, one black and the other white.
The 60 plants measured at each station all showed some terminal growth after 17 weeks. Those at the edge of the pond where the plant was the most abundant grew the least. Plants growing at the other six stations were indistinguishable, either statistically or by inspection. Temperatures during the experimental periods at the seven stations ranged from 29° to 40ºC, maximum; 15° to 23°C minimum. Maximum values fluctuated considerably between stations but minimums were very similar. Mean water loss from atmometer bulb varied from 50 cc to 235 cc. The climatic data obtained did not show any correlation with growth and distribution. Additional collections at selected stations at lower levels, however, indicated the detrimental influence of shading, a probable explanation of the poor growth at the heavily shaded edge of the pond.
On August 1, 1953, two samples of Spanish moss were taken from Putney Pond, another solution pond in the northeast part of Baker County. These samples were both taken from cypress trees, one at the edge of the pond and the other about 45 m away. The sites did not differ in the amount of wind and shading. Forty specimens from each sample were measured and analyzed in the same manner as those from Mossy Pond, that is, the mean, standard error, standard deviation, and coefficient of abmodality were computed for the growth data. Growth of the plants on these two trees was not statistically different nor were subjective differences evident.
Biotic enemies
No natural biotic enemies of Spanish moss were observed which affect it sufficiently to modify the range of distribution. The list of insects which have been found on the plant is extensive but does not indicate which are transients and which are permanently associated with the plant. Rainwater (1941) has found 163 species in the festoons, not including spiders. The caterpillar of the moth Dahana atripennis eats the young shoots, but this was observed at only two locations and in one experiment. These caterpillars ate the young Spanish moss tips placed with them in finger bowls, made their cocoons, and later emerged as moths. One bat (Lasiurias seminole) was found resting within a festoon of Spanish moss, an occurrence also reported by Hamilton (1943). The strands of the plant appear to offer a desirable place for insects to deposit their eggs judging from the number of eggs and egg cases which are found all over the festoons, particularly in the central portions.
Local distribution in relation to topography
Relationships between distribution and physiography were studied in an area about 12 miles west of the Emory University Field Station including portions of Early and Baker Counties. The area includes 156 ponds which have been described by Emory University Field Station personnel as permanent, semi-permanent, and temporary. The semi-permanent ponds contain some water except in the very dry seasons which may occur in the summer. The temporary ponds contain water only after a period of heavy precipitation and do not persist long after these rainy spells. The majority of the ponds have either mature pond cypress (Taxodium. disticum (L.) L. C. Rich.) or pond cypress in some stage of development at their edge. Only very few ponds are open with no trees. The greater portion of the area is under cultivation or has been in the recent past. The remainder contains mixed hardwoods and pines. There are two hardwood forests at Hilburn's Hummock and Big Cypress Pond which have not been greatly affected either by fire or lumbering. The ponds, as well as the rest of the area, were surveyed and classified in terms of the major trees, amount of water, and the presence (amount) or absence of Spanish moss.
With two exceptions, all Spanish moss was located on or near a pond. All permanent ponds which had Spanish moss in or around them contained a stand of mature pond cypress, while those ponds which had only immature cypress did not have any Spanish moss associated with them. When pines surrounded a pond, there was no spread of the epiphyte beyond the pines. In cases where both pines and hardwoods surrounded the ponds, a few hardwoods contained the plant and occasionally some were noted in a few of the pines. Generally, these were mature pines which had lost their characteristic pattern of dense foliage and had some exposed branches. Two adjacent ponds in the experimental area were separated by a stand of pines. One of the ponds contained mature cypress trees and the other only immature ones. Spanish moss was associated only with the pond having the mature cypress.
Distribution on individual trees
To obtain further data on distribution, three parallel transects 150 m long, 10 m wide, and 30 m apart were made across two sand hills and the intervening valley. The hills, near Newton, Georgia, parallel the Flint River and are approximately 400 m and 500 m from the river. Of the 65 trees examined, 40 had some Spanish moss in the crown. The trees included Quercus virginiana Mill., Q. velutina Lam., Q. Margaretta Ashe., Q. cinerea Michx., Q. falcata Michx.; and Crategus sp. The 25 trees containing no Spanish moss included the same oaks and one Pinus palustris Mill. There was no consistent difference in form, height, or other characteristics between trees having Spanish moss and those which did not, either between species, within a species, or between the two groups as a whole. However, here and elsewhere certain tendencies were evident. Very rarely does Spanish moss grow on the tips of live branches or twigs unless the amount of foliage is much reduced from the normal. Spanish moss tends to accumulate on the larger branches. When branches are exposed due to a break in the crown, the exposed portion tends to have the greatest amount of Spanish moss, especially on the periphery of a group of trees. The greater the amount of forking and branching, particularly on the smaller branches, the greater the number of small festoons and single plants.
In spite of the plant-pond relationship noted, many seemingly suitable ponds do not have any Spanish moss associated with them. Throughout the Coastal Plain, or even in one county, this epiphyte is spattered about in isolated populations often miles apart. The populations appear more haphazard and non-specific in their locations the further south they are located. This situation was apparent in central Florida where detailed records were kept on roadside distribution.
In the experimental area there was a definite correlation between the ponds and the occurrence of Spanish moss. Possible factors, other than those investigated, operating to control this distribution were observed at Big Cypress Swamp and Hilburn's Hummock, where there are large stands of virgin woods which have not been greatly disturbed by fire or lumbering. In these relatively undisturbed areas the Spanish moss extended to the limits of the trees surrounding the pond areas, while in the disturbed area the Spanish moss was limited to an area immediately surrounding the ponds. In a few cases a sufficient amount of arboreal reproduction had occurred since the last cutting or burning so that new hosts were available for the epiphytes. Spanish moss does not spread rapidly under normal conditions and all possible hosts had not been utilized in these areas.
It seems likely that the relationship between Spanish moss and the ponds and streams is due to the action of fire and lumbering which remove many of the suitable substrates for the plant. After the growth of new trees around a pond, Spanish moss spreads outward until stopped by some barrier. This barrier may be a treeless area or a thick stand of pines whose heavy foliage acts as a mechanical barrier. In addition, pines are not suitable substrates since the growing trees continually lose their lower branches, forcing the moss to somehow move upward to higher branches until the tree reaches maturity.
Still a third possible reason for the barrier offered by pines is the interception of rainfall. In the case of deciduous trees, more than 50% of the rain drops through the crown even in the weakest rainfall, whereas the needle-leaved species require a rainfall of at least 0.39 inches in order for half of it to reach the ground (Geiger 1950). The rain falling on the needle-leaved trees drops directly from the leaves, falling at the periphery of the trees. In addition, 20% of the rain falling on a deciduous beech runs down the trunk while less than 5% runs down the trunk on a needle leaved spruce (Geiger 1950). These trees are not involved in the present work, but some comparison may be allowed. Stem flow may be a major factor in the apparent preference of Spanish moss for broad-leaved rather than the needleleaved species. There is no toxic effect as evidenced by the number of seedlings and mature plants found on some of the older pines.
Many of the areas cleared for agricultural purposes lie adjacent to ponds and during cultivation do not offer any suitable sites for Spanish moss. Later, as these fields are abandoned, pines are usually the first arboreal invaders. Years later, if hardwoods replace these pines, there will again be an opportunity for Spanish moss to move out away from the pond areas. The long periods required for this succession and eventual emigration do not often occur in southwest Georgia. The prevailing agricultural practice is to burn over the old fields every winter, doing serious damage to the young pines, either killing them or retarding their development. These field fires often spread to the adjacent woods, damaging the trees and burning the festoons of Spanish moss which hang in the lower branches. `
The transect at Mossy Pond indicated that there is a definite vertical gradient within a given tree in that plants higher up show more terminal elongation than the lower ones. This distribution may be the result of shading. Other than this, there does not appear to be any specific pattern within a single tree. The dispersion does not show any consistency either within a tree or between species except that which results from the mechanical defense of pines and young cypresses, which resemble pines at that stage:
DISTRIBUTION
United States
Spanish moss is restricted to the Western Hemisphere, apparently having originated in South America and then spread eventually to North America (Smith 1934). All the counties in the United States in which Spanish moss has been recorded are either below or near the Coastal Plain fall line (Fig. 1). This is a topographic line and it seemed probable that some factor, or combination of factors, particularly climatic, which correspond to the fall line, might be the actual limiting barrier. The single climatic factor which best fits the northward limits of Spanish moss is the average vapor pressure of .425 inches of Hg (Meyer 1942). The line indicating an average annual temperature of 65º F is somewhat parallel and using this temperature to convert from mean average vapor pressure, the mean annual relative humidity is 63% (Fig. 12). Such a generalized figure, however, is of dubious value.
Spanish moss has successfully survived for 5 years in the mountainous region of northeast Georgia, at Toccoa. Plants were taken there as a curiosity by one of the residents and placed in trees in his yard. These plants have flowered, but there is no record of seed production. For this investigation, additional plants were sent to Toccoa and at least one viable seed was produced by these plants in season. Toccoa had an average annual rainfall of 65.5 inches over the 5-year period 1949-54, and the average temperature over this period was 62.3° F. The approximate minimum temperatures endured by the plants ranged from. 28° F in 1950 to 15° F in 1951. The trees on which these plants were located are sheltered and thus the United States Weather Bureau readings only approximate the actual temperatures affecting the plants. The minimum average monthly temperature over this period was 41.6° F in December of 1951 while the maximum average monthly temperature was 81.5° F. The low temperatures are generally lower than those occurring within the boundary limits of Spanish moss in its present United States distribution.
Along the Gulf of Mexico in the states of Florida and Alabama, Spanish moss occurs in sporadic populations which are separated by long distances. There is no radical change in the major trees and vegetation along the coastal Highway U. S. 90, near Biloxi and Gulfport, Mississippi, but there are several gaps of 5 and 6 miles between populations. This discontinuous distribution is also true inland. These isolated areas containing Spanish moss are not necessarily associated with ponds and may be dry, sandy oak barrens. A wooded area 23/4 miles north of Ellaville, Georgia, contains a heavy concentration of Spanish moss confined to a small portion of the total wooded area. The nearest known population to this area is nearly 22 miles to the south. Such situations are common throughout all of the states investigated.
Central and South America
Because of the peculiar distribution of this plant in the United States, the distribution of Spanish moss in Central and South America was surveyed using herbarium specimens from the Smithsonian Institute National Herbarium. Of the 173 sheets examined, 41 were located and the climate was approximated (U. S. Department of Agriculture 1941). In a few cases, the local conditions were recorded by the collectors.
In the United States Spanish moss does not grow much above a 33-meter (100-ft) elevation. This height distribution has been cited as proof of the plant's dependency on some "unknown" geophysical factor, or some related condition, with relative humidity being the one most often indicated. Plants were found to occur normally at about 3,000 m (10,000 ft) m South America. Several specimens were taken in the vicinity of La Paz, Bolivia, at altitudes from 2,600 to 3,000 m (7,800 to 10,000 ft). The maximum temperature there is 75° F and the minimum is 27° F in the city. One may assume that the higher altitudes have a more severe climate. The annual precipitation is 22.18 inches, below the average of most of our Coastal Plain regions. Cool temperatures apparently are not required at any time in the life cycle since the plant grows well at Merida and Cuidad Bolivar in Venezuela. Ciudad Bolivar has a temperature range from 66° to 97º F with an average annual precipitation of 35.15 inches. Merida ranges from 52° to 85° F with an average precipitation of 71.83 inches.
Spanish moss survives extreme temperature fluctuations from 25° to 109º F in Montivedo, Uruguay. It survives under very low rainfall conditions as shown by its growth in Lima, Peru. The annual rainfall there averages 1.9 inches and the temperature ranges from 40º to 90º F.
The explanation which seems most likely for the present distribution of Spanish moss in the United States is that the plant is entirely windborne whenever it moves at any stage in its life cycle. Van Cleef (1908) has indicated the principal average storm paths moving across the United States, based on 1,600 storms. The southern paths of these cyclonic storms could well account for the main lines of the present distribution. Smith (1934) gave evidence to support the view that Spanish moss had its origin in South America and from there moved northward to Central America and Mexico as well as to many of the islands of the Caribbean. Once the plant reached northern Mexico, the storms as depicted by Van Cleef could have carried the plant over most of its range in North America. Some portions of the original stock may have come into this area from the islands south of Florida in the Greater Antilles by means of hurricane paths as shown by Tannerhill (1952). The main paths of both of these storm groups are shown on Fig. 12. Certainly some of the present distribution of Spanish moss is a result of local influences as described above, but these factors have probably not played a very important part in the extension or limitation of the overall range. Instead the present range and distribution of Spanish moss may be the result of primarily mechanical factors and would therefore be entirely fortuitous. It seems that all of the climatic conditions which are required for the successful establishment of Spanish moss are met within the entire southeastern Coastal Plain.
Plants examined from different portions of nine states, Central America, and South America did not show any consistent morphologic differences, possibly indicating a relatively stable genetic complex.
The factors which prevent the plant from migrating out of the Coastal Plain are not clear, but temperature in itself does not seem to be limiting. The plants at Toccoa, Georgia, have survived for 5 years under relatively severe conditions. In two cases, plants found above the fall line in Georgia are in depressions cut by streams (W. H. Duncan, private communication, 1953). As was shown earlier, there is a partial correlation between the fall line and a relative humidity of 63%, with this value increasing toward the ocean. It may be that above the fall line the lower relative humidity coupled with less frequent rains is insufficient to sustain the plant. It is extremely difficult to correlate vegetational patterns with isopleths. There are undoubtedly areas above the fall line more suitable to the plant than many of the locations where it now grows.
- Hoaxes and non-hoaxes by Butcher
In 1998 I was given a test by Tom Lineham( past Editor of the BSI Journal) who sent me a photograph entitled Tillandsia usneoides but the photo had been taken somewhere in Alaska. If you did not know the locality you would assume it was correct like I nearly did. It is in fact a Usnea.
In 2002 I was sent a photo of Racinaea insularis by Sandra Pozo on the Galapagos Islands. But what were the hanging festoons in the photo? Was this a first sighting of T. usneoides on these islands. False alarm because they were lichen of some sort!
In 2006 on my asking on the internet on Brom-L about a blue flowered T. usneoides as mentioned in Smith & Downs, Eric Gouda of the Utrecht University, Belgium obliged by doctoring a photo so we had a blue petalled flower. He even sent a red flowered one for good measure. By this time even the most naïve were having doubts of these claims!
We must remember that the petals of this Tillandsia are not always green. In 2000 when Wolfgang Tittelbach - Editor of Die Bromelie - said there was a yellow flowering form, I took it with a pinch of salt! Anyway, I acquired a small piece and a couple of years later it flowered yellow which had me in raptures and bragging to Walter Till. Walter deflated me somewhat, by saying it was found somewhere in N Peru and was not that rare. I decided it should be recorded somewhere and this is how the cultivar name 'Spanish Gold' came into existence. The interesting thing is that it sets seed readily but the seed is not viable. Perhaps it needs another clone.
This seed setting had me wondering because although I was always seeing green flowers on my various bunches of T. usneoides hanging around the garden I had never seen any seed. This puzzled me because of my experience with the sub-genus Diaphoranthema of which T. usneoides is a member. I have rather an extensive collection of this group and find that if a plant looks like it is going to flower you always seem to find a seed pod later on. On some occasions the flower does not even need to open! Why is it so?
In 2006 Greg Dauss of California sent me a photo of roots on a T. usneoides seedling. I knew this was not a hoax because I knew that you could not get roots on a 'cutting', seedlings were still 'smart' enough to anchor themselves wherever they germinated by the use of roots. I just had to ask Greg how many times was he getting seed to set on his T. usneoides. In his experience he had only seen the thick leaved form perform.
Perhaps others may like to comment on their experience.
Any more hoaxes or non-hoaxes?
- Varieties in Mez 1935
Var. filiformis Andre, Bromel. Andr. (l889) 64. - Habitu tenuissimo, foliis internodiisque elongatis, vix ultra 0,5 mm diam. metientibus - .Habit very slender, leaf internodes long, barely more than 0.5mm diam.
Colombia (Andre), Guyana (Schomburgk n. 159, Wullschlugel n. 1099), Brasilien (Gaudichaud n. 134, Regnell III. n. 1251, Schwacke n. 6454, Ule n. 647, Wawra I. n. 890), Bolivia (Mandon n. l182), Paraguay (Balansa n. 611).
Var. ferruginea Andre, l. c. - Habitu tenuissimo sed internodiia saepissime abbreviatis; foliis vix ultra 20 mm longis.- Habit very slender but very often internodes shortened, leaves barely more than 20mm long
Bolivia (Bang n. 107, Hauthal n. 351, Hoffmann). Paraguay (Tweedie n. 526). Argentina: here the most frequent form (Castellanos n. 1197, 25/1857, 27/2346, Hieronymus n..858, 358, 455, Haumann. 106, Kurtz n 9828. Lorentz n. 797, Lorentz und Hieronymus n. 449, Stuckert n. 11322). Chile (Bertero, Cuming n. 848, Reiche)..
Var. longissima Andre, l. c: - Internodiis foliisque (his saepius ultra 0,1 m metientibus) elongatis crassiusculisque (his ad 1 mm crassis). - leaf internodes (often more than 10cm long) elongated thick ( to 1mm thick)
Colombia (Kalbreyer n.332). Venezuela (Fendler n. 1535). Trinidad (Fendler n: 818). Brasilien (Glaziou n. l225).
Var. robusta Morr. sp. Mez in Mart. Fl. Brasil. III. 8. (1894) 613. - T. usneoides ? major Andre, I. c. - Habitu compacto, internodiis abbreviatis, foliis l,5-2 mm crassis. - Compact habit, short internodes, leaves 1.5 - 2 mm thick
Mexico: not rare (Arsene n. 466, 848, Aschenborn n. 196, Berlandier n. 460, Bourgeau n. 95, Coulter n. 1578, Galeotti n. 48i8, Karwinsky n. !4, Palmer n. 124, Purpus n. 3395, Schaffner n. 20,225, Schiede n. !002, Wawra I. n. 401 e. p.). Peru (Dombey n. 161, Mathews n. 652).
Var. cretacea Mez, I. c.:- Praecedenti simillima, sed differt habitu perconglomerato, foliis caulibusque magis quoque niveo- vel cretaceo-lepidotis. - Very similar to above, but differs by clumping habit, leaf stem larger each white or grey lepidote
Peru: one collection (Philippi).
- A STUDY OF TILLANDSIA USNEOIDES.
by Frederick Billings in Bot. Gaz. 38: 99-121. 1904.
(WITH ONE FIGURE AND PLATES VIII-XI)
Tillandsia usneoides, popularly called "long moss," "black moss," -or "Spanish moss," is the most widely distributed representative of the tropical and subtropical family Bromeliaceae. According to SCHIMPER (I) it extends from southern Virginia, its northern limit, as far southward as the Argentine Confederation. It forms everywhere a conspicuous and characteristic object of the landscape, its long gray festoons adorning not only trees of the virgin forest but many cultivated ones as well. Although the beauty of the landscape is enhanced by its presence, its growth upon ornamental trees is regarded often with apprehension, a common impression being that it lives parasitically. A most casual examination, however, will reveal the fact that the moss is in no way connected with the tree, but merely wraps its dead, wiry stems loosely around the twigs in order to support itself. Old festoons which have hung in the same place for years occasionally show a connection with the bark, the annual growths of the limb finally enclosing some of the decorticated moss stems; much in the same way that an old horseshoe hung astride a branch and left unmoved for a long time will be partially enclosed.
An indirect cause of the popular belief in the parasitism of Tillandsia is its preference for sunny exposures. This habit would tend to keep it from trees having a dense shade. In dark forests it hangs suspended from the higher limbs of tall trees, especially those that are dead. Many a cultivated tree when in perfectly healthy condition possesses too dense foliage to serve as a host for Tillandsia, but if for some reason the supply of leaves should be reduced, the light conditions might be such as to make the presence of the. epiphyte possible. Should it make its appearance, the owner of the tree would be very apt to regard the moss as the cause rather than the result of the reduced foliage. A proof of the true epiphytism of the plant is its long continued and vigorous growth upon decorticated limbs of dead trees. Near Baton Rouge are many such trees, killed by girdling long ago, yet supporting a large quantity of moss. In order to demonstrate experimentally that the moss can live solely on what it derives from air and rain, some festoons were supported by twine and hung from some branches of a tree upon which moss was already growing. As was expected, the festoons produced normal flowers, gave rise to new growth, and at the end of eighteen months looked as vigorous as any on the tree, though they came at no time in contact with it.
Because Tillandsia has no influence as a parasite, it does not follow that it exerts none in other ways, yet to just what extent it affects a host tree is at present difficult to say. Aside from the slight damage done in breaking twigs and small branches by its weight, it is doubtful whether such objections as shading and cutting off the supply of air are really worthy of consideration. It is almost certain that these objections are not sufficient to explain a reduction in foliage that people so often ascribe to the presence of the moss. It is realized, however, that this problem can only be answered satisfactorily by experiments extending over a considerable number of years.
The problem of the distribution of T. usneoides upon the various species of trees is one of the first to force itself upon the observation. That certain trees of a given locality are abundantly supplied while others not far distant are not, is a well-known fact. One factor in the case has already been mentioned, and that is the light relation. But there are others to be considered, and the most important perhaps is concerned with the method of dissemination. The epiphyte is not usually propagated by seeds but by fragments of festoons, which being somewhat heavy cannot be carried far except in a very high wind, or by birds, which according to SCHIMPER (1) in some regions utilize the plant in building their nests. There is a good chance, therefore, for a tree a little distant from others bearing the moss not to receive its first detachment of the epiphyte.
The character of the foliage also plays a part, in that a tree with leaves densely crowded on the outermost twigs would scarcely permit a wind-blown fragment of moss to hook itself to the branches, but would shed it. SCHIMPER (I) observes in this connection that "Baume mit sehr dichtem Haube entbehren der Sonnenepiphyten beinahe ganzlich." According to PEIRCE (2) Ramalina reticulata, a lichen having a habit and mode of dissemination. similar to T. usneoides, is found more frequently on deciduous than on evergreen trees, because, as he explains, the foliage of the evergreen trees interferes with its reaching the branches. The umbrella tree (Melia Azederach) has a remarkably dense foliage and is almost universally devoid of moss, yet near the university is a tree of this species with a scanty supply of foliage and an abundance of moss. It is reasonable to conclude that any tree furnishing proper conditions for attachment and growth may become a host of the epiphyte.
The source of the water supply of Tillandsia is atmospheric precipitation, as in all epiphytes. Dissolved in the water are the necessary salts which have been dissolved by the rain from the dust in the air. Perhaps an equally fruitful source of salts is in many cases the washings from the tree, which in dry weather may accumulate much earthy material in the form of dust upon its branches. The plant itself even serves in collecting dust on account of the scaly surface, so that when wet the deposits beneath the scales yield a small amount of soluble material.
A most remarkable characteristic of Tillandsia is its ability to retain water. The absorption of water is accomplished over the entire surface of the living parts by means of scales, as will be described further on, its retention being accomplished also by the scales, and of course by the cuticularized epidermis. It is much easier to understand how a melon cactus with its globose form and consequent minimum surface and enormously developed water-storage tissue can resist prolonged drouth than it is to see how Tillandsia with its small cylindrical leaves, much greater surface exposure, and comparatively small storage facility can, without any water supply, endure drought. A small festoon was hung in a closed dry room for nineteen days without water. It lost 23 per cent. in weight during the time, but when placed in water it absorbed as much as it had lost, and remained a healthy plant, showing that it had not really suffered injury by exposure to the drought. There is occasionally, of course, a similar drying process in the open air when drought occurs. During the dry spell in the spring of 1902, moss plants were known to have been subjected to two months of rainless exposure without injury.
From an economic standpoint, Tillandsia is of some commercial value on account of its mechanical tissue. This forms a central cylindrical strand composed of reduced phloem and xylem, surrounded by a mass of thick-walled sclerenchyma fibers. When the parenchymatous cortex is removed, the sclerenchymatous axis remains as a tough elastic fiber, which serves as a packing in upholstery. The so-called curing process is a means of eliminating the parenchyma. One method largely employed is that of burying the moss in trenches or pits, allowing it to remain till the cortex is dead and in a condition to be removed easily.
DEVELOPMENT OF THE EMBRYO SAC.
The primordia of the ovules arise on the innermost wall of each loculus of the tricarpellate, superior ovary. By a one-sided growth each primordium becomes bent toward the base of the ovary, developing into the anatropous type of ovule. When the bending has reached an angle of about 90º, the nucellus appears as a hemispherical mass of cells, at the base of which can be seen the beginning of the inner integument. Imbedded under two layers of nucellus cells, the single archesporial cell becomes differentiated in the usual way, by its slightly larger size and greater staining capacity (fig. a). As the ovule increases in size, the nucellus elongates, the outer integument appears, and the archesporial cell enlarges considerably, especially in length. There is no parietal cell formed, but by multiplication of cells the nucellus over the archesporial cell forms an additional layer, making three (fig. 3). The archesporial cell is now much elongated, and occupies the central region of the nucellus. It is filled with granular, longitudinally-striated cytoplasm, and has a relatively large nucleus. The first and second divisions of this nucleus probably give rise to the gametophyte generation. Only one spindle of the first division was observed, and it was but little more than one-third the length of the cell (fig. 4). The chromosomes were short, and closely crowded at the equatorial plate. The conditions were altogether unfavorable for ascertaining their number on account of the small size of the figure. The number, however, was definitely made out from the second division of the pollen mother cells, and was found to be sixteen. A protracted search failed to yield a nuclear figure which definitely showed the chromosome number in the sporophyte, though considerably over sixteen were observed.
The first division of the archesporial cell is usually followed by a transverse wall and a resting condition of the nuclei (fig. 5) ; but a single case was observed, as reported by SMITH (3) for Eichhornia crassipes, in which a row of four nuclei was formed without separating walls (fig. 7). In Eichhornia the absence of the walls is said to be the rule, but in Tillandsia it is the exception. The division which gives rise to the third and fourth megaspores, thus completing the axial row, will be seen from fig. 6 to be in the cell nearest the micropyle. In the meantime, the basal of the two proximal megaspores begins to elongate, and is destined to develop into the embryo sac. A vacuole is formed in this cell as it pushes outwards crushing the other three megaspores, whose contents soon show evidence of breaking down. The remaining stages in development are the familiar ones of complete absorption of the non-functional megaspores by the functional, and the internal division of the latter into eight cells.
The two cells that are to form the synergids soon come to possess larger nuclei than does the egg cell. The egg nucleus in fact is smaller than is customarily observed. In the completed embryo sac, the egg often lies against the wall of the sac near one synergid, but may occupy a position between the synergids. The polar nuclei usually approach each other and fuse near the antipodal region (fig. 14). The antipodals occupy a pocket at the extreme end of the sac.
FERTILIZATION.
The pollen tube passes through the micropyle, penetrates the nucellus, and enlarges as it enters the embryo sac. It does not appear to pass between the synergids, but to one side of them, one synergid being disorganized in the process. The two male nuclei which have arisen from the generative nucleus during the development of the pollen tube lie near together and a little in advance of the tube nucleus. In no case observed did the male nuclei show the much elongated, spermatozoid-like form so often described for other plants. In fig. 15, which represents the tube before its rupture, they are elliptical; but when discharged they are slightly more elongated and may have pointed ends. The place of discharge may be either at the end of the tube or lateral, though near the end ( figs. 16-19). The tube nucleus is usually to be seen at the time of discharge of the male nuclei, but may be absent later, which would indicate that it too was ejected. In one instance (fig. 19) the nucleus was observed after ejectment. The male nuclei are of about the same size and appearance, and leave the pollen tube at about the same time.
The nucleus which is to fuse with the endosperm nucleus can be seen in various stages of its passage to the antipodal end of the embryo sac. There is no evidence that either nucleus increases in size after leaving the pollen tube. The time of fusion with the polars may be either before or after their complete union with each other; in fig. 18 it is before. In fig. 18 the fusion of the two male nuclei with the egg and polar nuclei respectively is seen to be simultaneous. After fertilization the egg secretes a wall about itself and rests for a time.
The occurrence of darkly stained bodies so frequently seen in pollen tubes has been noted in Tillandsia. They were observed in the microspores before germination, which would account for their presence in the pollen tube.
THE SEED.
The most noticeable change that results from fertilization is the extensive elongation of the entire ovule. Part of the growth is due to enlargement of the embryo sac and its surrounding integuments, while the remainder is traceable to elongation of that part of the outer integument which is prolonged above the body of the ovule. The inner integument does not appear to elongate at all, hence the opening of the micropylar canal comes to lie far below the opening of the canal formed by the outer integument (fig. 22). A similar elongation of the outer integument was observed in Puya chilensis by HOFMEISTER (4).
Accompanying the growth of the embryo sac is the development of the endosperm. It begins to form at once after fertilization, and the nuclei resulting from the first divisions of the endosperm nucleus take position at either end of the sac, leaving, however, a few to form
a thin parietal layer between. At the antipodal end, cell formation with walls begins at once, and a number of large cells form a tissue which stands out conspicuously in the cavity of the sac, which otherwise contains only a few free endosperm nuclei. At first this tissue was taken as an extraordinary development of antipodals, but cases were found where the three degenerate cells were lying beneath the tissue in the small pocket at the end of the embryo sac. The free endosperm nuclei gradually gather in increasing numbers against
the endosperm tissue, finally forming walls about themselves but remaining readily distinguishable from the other tissue (fig. 24). The functions of the two tissues appear to be somewhat different. The originally formed cell-compact retains its richness of protoplasmic contents during the development of the embryo, probably serving in the conduction of food materials to the later formed tissue adjoining it, which soon shows signs of containing food deposits. The reserve materials thus laid down are not utilized by the embryo before seed germination, but exist as the endosperm of the ripe seed. The endosperm at the micropylar end of the embryo sac does not develop in large quantity, forming a tissue about the embryo only after the latter attains a considerable size.
The egg cell remains dormant for a time after fertilization. In 1903 the period of blossoming lasted (at Baton Rouge) for a month following the middle of May. Material gathered about the first of July showed egg cells undivided, as well as embryos of only a few cells. Growth during the summer is slow, small embryos being found in material gathered about the tenth of August. It was not till the middle of September that large ones were observed, and even then there was much diversity in size.
The first wall formed in the division of the egg cell is transverse, as is the second one also. The proembryo of three superimposed cells is therefore not different from the type that holds in so many monocotyledons. The divisions immediately following, however, vary considerably in sequence.
The middle segment may divide sooner than the terminal (fig. 28), or the reverse may be true (fig. 27). The basal segment divides sooner or later by longitudinal walls into four cells - a variation from the Alisma-type, in which the segment is unicellular and vesicular. The terminal segment divides by longitudinal walls to form the quadrant, and by transverse walls to form the octant. The latter walls instead of being precisely transverse may be oblique (fig. 34). In many older embryos the arrangement of the cells in this segment indicates that the walls in question were originally oblique or else became so by unequal growth in different parts of the embryo (fig. 36). The dermatogen usually forms first in the terminal segment. To distinguish the middle from the terminal segment soon becomes a difficult matter, but from the position of the concavity in which the stem apex is developed, it is safe to say that the apex arises from the middle and the cotyledon from the terminal segment, as in Alisma. The middle segment also gives rise to the root-tip, hypocotyl, and part of the suspensor. A short time before the differentiation of the stem tip in the lateral depression, the region adjoining and outside of the area where the stem tip is to appear grows upward into a ridge of tissue, which in the mature embryo encloses the growing point completely. If the figure of the embryo of Guzmania, as shown by WITTMACK in Engler and Prantl's Naturlichen Pflanzenfamilien be compared with that of T. usneoides (fig. 40), the resemblance will at once be apparent. It will be noticed that what I have called cotyledon in Tillandsia is called scutellum in Guzmania, the term cotyledon ¹ being reserved by WITTMACK for the small outgrowth labeled c, near the stem apex. It is probable that the author in thus naming the two organs scutellum and cotyledon only wished to emphasize the difference in function, one as an organ of absorption, the other as a rudimentary leaf, at the same time recognizing the two as homologous with the cotyledons of the dicotyledons. From a study of the seed germination of T. usneoides, however, it will be seen that it is extremely doubtful if the organ named cotyledon in Guzmania is really such. Further discussion of this point, however, will be postponed till seed germination is considered.
When the embryo of Tillandsia is about three-fourths grown, there occurs a degradation of certain cortical cells of either the root or the end of the hypocotyl nearest the root-tip. The cells in question show at first a contracted protoplast, with incapacity to stain deeply, and by the time the embryo has reached its full size almost a complete absence of cell contents (fig. 42). This phenomenon undoubtedly stands in intimate relation with the complete atrophy of root that obtains in the mature plant.
_____________________________________________________________________
¹The index letter c in the description of fig. 19, G of the Bromeliaceae has been found through correspondence to indicate cotyledon.
Dispersal of seeds in the Tillandsia is accomplished by the assistance of long delicate hairs that beset the seed coat. These arise by elongation of the cells of that part of the outer integument which forms a portion of the body of the seed, and also from that part which extends to the funiculus. The hairs not only assist in wind transportation, but are also of use to the seed in enabling it to adhere to bark or festoons of moss. The adaptation for effective adherence consists in closely appressed barbs attached to the hairs at intervals (fig. 44). Soon after the opening of the capsules, numerous instances of seeds clinging tightly to limbs and to moss festoons may be observed.
The time of discharge of seeds is in March (at Baton Rouge). I have no data as to possible variation of this time in localities widely distant, but suppose it is nearly uniform for the southern states. March, of course, is an unusual month for dehiscence of fruits in the north temperate zone, but in Tillandsia it stands in close relation to another property not generally possessed by seeds in temperate climates, that is, quick germination. Though lack of facts forbids positive statement, it may be conjectured that this relationship originated from ancestors living in tropical lowlands, where a dormant period to withstand unfavorable conditions is unnecessary.
GERMINATION OF THE SEED.
Tillandsia produces seed in considerable quantity each year. Just what proportion contains fully-matured embryos has not been ascertained, but there is no doubt that a large percentage have them. The embryos appear perfectly normal, with the exception of the dead cortical cells in the root or hypoctyl, and show no apparent reason why they should not give rise to seedlings. The experience of investigators, however, has been that seeds produced by the epiphyte are worthless, a condition which has arisen through the introduction of a vegetative mode of reproduction, whereby seed-production has degenerated. Nevertheless, I made efforts to induce seeds to germinate by placing them in a germinator, but without success. MEEHAN (5) reports having found the seed germinating in the hollow- crotch of a tree in which vegetable mold had collected. He says that from the seedlings or young plants proceed stolons or runners, having buds every few inches, which push out into leaves and stems to form the gray-green moss. SCHIMPER (1) succeeded in finding one seedling, but he gives no description of it.
MEZ (6) states he was unable to obtain any seedling at all. Realizing that the observations of MEEHAN were worth consideration, I searched crotches of moss-laden trees, in which plenty of vegetable mold had collected, but without success.
I then planted seeds in the mold, but they could not be induced to germinate. On April 6, 1903, I observed Tillandsia seedlings for the first time, and they were projecting from a partially-opened capsule (fig. 1). Out of the nineteen seeds in the capsule, thirteen had developed into seedlings. They were held in place by the tuft of hairs from the testa to which they still adhered. An examination of moss festoons was then made, with the result that many little seedlings were found either still attached to the capsules, or else hanging to the scaly stems and leaves of the mother plants. In every case the seed coat still adhered to the base, or root-end of the seedlings, so as to enable the coma to keep them from falling to the ground, which they certainly would have done without this provision. When it is remembered that the capsules dehisced in March, and the seedlings were found early in April, it will be seen that germination followed dehiscence quite closely. Of course the early growth was attained at the expense of the endosperm, but when it was exhausted, continued growth, which would naturally be expected from healthy looking seedlings, failed to occur. Material gathered in the summer and autumn yielded the usual crop of seedlings, but in no case were any found that were larger than those found in April. Festoons gathered the middle of January, nearly a year after the capsules opened, had numerous little seedlings hanging to them, all healthy looking, but no larger than any observed before them. It is expected that when the warm weather of spring comes, when Tillandsia puts forth its most vigorous growth, the seedlings also will increase in size. The question naturally arises here, why Tillandsia seedlings are not to be seen in all stages developing into mature plants, counting of course those which germinated previous years. As such is not the case, it can only be conjectured that, as the spring of 1903 was an unusually rainy one, the conditions for germination were especially favorable.
Seedlings exhibiting various stages in germination were imbedded in paraffin and longitudinally sectioned. In the earliest stage (fig. 45) the first leaf shows only a slight growth, the stem apex is still undifferentiated, while from the axil of the ridge of tissue that enclosed the stem apex, or else from its inner surface, a pair of organs have arisen. It is believed that the presence of these organs throws some light upon the morphological nature of the ridge of tissue. If a section is made through the nodal region of a mature plant (fig. 49), it will be seen that the leaf sheath which encloses the lateral shoot and main axis is double. The doubling is not due to splitting of a tissue once entire; but to bifurcation. A section through a very young sheath (fig. 49a) reveals an outgrowth, one to several cells in extent, from which a double layer of cells arises. These soon separate to form the double sheath. In older stages the base of the sheath is composed of many cells in width, so that the sheath appears no longer to originate as a bifurcation of a single organ, but rather as two distinct organs. Both organs or portions of the sheath may develop equally, though it more often happens that one portion becomes larger than the other. Occasionally, the inner scarcely develops at all, but remains a tiny rudiment.
The sheaths which arise in the seedling develop precisely like those in the mature plant and differ from them in no respect. The two organs that originate on the ridge of tissue, therefore, may be regarded without hesitation as the first sheath, and as every sheath appears in connection with a leaf, that leaf must be the cotyledon. From the section of the mature plant it will be noticed that the bases of each leaf and its sheath are at the same level on the axis. If a difference in level should occur, however, whereby the base of the sheath were elevated above that of the corresponding leaf, the cell growth producing that elevation would originate from the cortical parenchyma lying immediately under the sheath. The parenchyma would give rise to a ridge bearing the sheath upon its summit. Such an occurrence does not of course actually take place in the mature plant, but it is believed that it is in such a way that the ridge of tissue originates in the embryo. Reasons for coming to this conclusion are based upon
the position of the first sheath. While the inner portion of the sheath may grow from the crotch at the base of the ridge of tissue, the outer, and sometimes the inner also, is attached to the ridge upon its inner surface. The outer portion may in fact arise from the summit of the ridge. The base of the sheath, therefore, is on the whole raised above that of the cotyledon, the elevation being accomplished through growth of the subjacent parenchyma. Thus there develops a special organ which serves a special purpose, perhaps as protection to the stem apex, and which must therefore be regarded as an embryonic structure without an exact counterpart in the adult plant. It cannot be a leaf, or a cotyledon, because a leaf does not bear such a relation to its sheath. A leaf and its sheath always develop with a growing point between them, so that they can never join in a median section. CAMPBELL (7) calls a similarly placed though less extensive outgrowth in the embryo of Sparganium a sheath. While it does not require a stretch of the imagination to consider the growth in question a sheath, there is at least one objection to this solution of the problem. The development of the sheath shows that it appears as a bifurcated organ almost from its incipiency, and that the base, at first narrow, subsequently increases greatly in width. Quite the reverse would be true in the embryo if the organ enclosing the growing point were regarded as a sheath, for the basal portion is first enormously developed, leaving the upper bifurcated portion to appear comparatively late.
The stages in germination are shown in figs. 45-48, which should be compared with fig. 49. The latter exhibits a difference in relative time of differentiation of stem and leaf apex as compared with the seedling. In the mature plant the leaf is still quite small when the stem apex becomes distinguishable at its base, while in the seedling the leaf first attains considerable size.
THE FLOWER.
The flowers, which are produced in considerable quantity in May and June, present little of special interest. Each flower has a calyx of three sepals, and a corolla of three green petals. Having a fragrant odor, it is possible that it is visited by insects, though no information has been collected by me on the subject. Thrips, however, inhabit many of the flowers and puncture the style in order to deposit an egg at its base. It is possible, therefore, that they may serve in cross pollination.
Although the flower appears to be terminal, it is regarded by MEZ (6) as a reduced indeterminate inflorescence. An examination of preparations made longitudinally through buds bears him out in his statement, for a growing point of considerable size is present, though having dead meristem tissue.
THE LEAVES.
The leaves of T. usneoides are acicular and with an approximately semicircular cross section. The epidermal cells do not have specially heavy walls, nor are the inner ones thicker than the outer, as in certain other Bromeliaceae. Sections through the leaf show it to have three fibrovascular bundles, each surrounded by a tissue composed of thick-walled sclerenchyma fibers (figs. 50, 51). The principal portion of the leaf is composed of parenchyma cells which do not show any differentiatiation at all into palisade and spongy tissue. While the cells have the shape of those in typical spongy tissue, the large intercellular air spaces characteristic of most mesophytic leaves are here replaced by small ones, giving the whole tissue a much more compact appearance. Not all of the parenchyma cells contain chloroplasts, for there are interspersed cells without them, whose function is that of water-storage, having walls provided with large pits which facilitate the passage of water from one cell to another.
Aside from acting in the capacity of mechanical tissue, the vascular system has undergone a process of degeneration. The necessity for a functional xylem with its transpiration stream is eliminated by the fact that there is a complete absence of roots, and also by the fact that the water-absorbing organs, the scales, are found over the entire exposed surface with the exception of some of the floral organs. There would appear also to be no need for a functional phloem since all living cells either contain chlorophyll and are exposed to light, or else are approximate to those containing chlorophyll.
THE CHLOROPLASTS.
One of the most interesting features of the leaf is the structure and behavior of the chloroplasts. These bodies, instead of exhibiting the more or less homogeneous structure observed in most chloroplasts, are seen to be composed of masses of smaller chloroplasts, measuring about 2 µ long and about a third as wide (fig. 52). While a very few cells in every cross section of the living leaf contain chloroplasts of the usual type, the vast majority of them contain such as have been described above. The little chlorophyll bodies have almost, if not quite, the minuteness of bacteria, and for convenience will be spoken of as microchloroplasts; the larger bodies, of which they appear to form a part, being distinguished as megachloroplasts. The true significance of the formation of the microchloroplasts will be readily seen when it is stated that they may not remain in bunches (fig. 52), but can and often do separate from one another till the entire cytoplasm of the cell becomes dotted with them (fig. 53). Under a low magnification such a cell appears uniformly green throughout. They even enter the vacuoles, where a lively Brownian movement is set up.
of the leaf is composed of parenchyma cells which do not show any differentiatiation at all into palisade and spongy tissue. While the cells have the shape of those in typical spongy tissue, the large intercellular air spaces characteristic of most mesophytic leaves are here replaced by small ones, giving the whole tissue a much more compact appearance. Not all of the parenchyma cells contain chloroplasts, for there are interspersed cells without them, whose function is that of water-storage, having walls provided with large pits which facilitate the passage of water from one cell to another.
Aside from acting in the capacity of mechanical tissue, the vascular system has undergone a process of degeneration. The necessity for a functional xylem with its transpiration stream is eliminated by the fact that there is a complete absence of roots, and also by the fact that the water-absorbing organs, the scales, are found over the entire exposed surface with the exception of some of the floral organs. There would appear also to be no need for a functional phloem since all living cells either contain chlorophyll and are exposed to light, or else are approximate to those containing chlorophyll.
THE CHLOROPLASTS.
One of the most interesting features of the leaf is the structure and behavior of the chloroplasts. These bodies, instead of exhibiting the more or less homogeneous structure observed in most chloroplasts, are seen to be composed of masses of smaller chloroplasts, measuring about 2 µ long and about a third as wide (fig. 52). While a very few cells in every cross section of the living leaf contain chloroplasts of the usual type, the vast majority of them contain such as have been described above. The little chlorophyll bodies have almost, if not quite, the minuteness of bacteria, and for convenience will be spoken of as microchloroplasts; the larger bodies, of which they appear to form a part, being distinguished as megachloroplasts. The true significance of the formation of the microchloroplasts will be readily seen when it is stated that they may not remain in bunches (fig. 52), but can and often do separate from one another till the entire cytoplasm of the cell becomes dotted with them (fig. 53). Under a low magnification such a cell appears uniformly green throughout. They even enter the vacuoles, where a lively Brownian movement is set up.
It is offered in explanation of this interesting condition of affairs that the supply of light of Tillandsia is considerably diminished by the presence of the overlapping scales, which are necessary for water absorption and for protection against too rapid transpiration. In order to meet this diminution, it not only prefers sunny exposures, but has modified its chlorophyll-bearing apparatus by causing it to occupy a much larger area in order to utilize to better advantage such light as penetrates to the interior of the leaf. .
It may be stated here that precautions were taken to examine healthy festoons removed directly from moss-laden trees. In some instances these were examined immediately after such removal, lest confinement in the laboratory should in some way induce pathological conditions.
THE SCALES.
The scales cover the entire living exposed portion of the plant with the exception of the corolla, stamens, ovary, and a portion of the calyx. Each scale develops from a single epidermal cell, the early divisions of which occur while the young leaves and stems are included within the leaf sheath. The first division is transverse (fig. 55). The proximal cell thus produced remains undivided, the distal dividing transversely till four cells are produced, of which the lower three form the stalk of the scale (fig. 57). The outermost hemispherical cell becomes divided into four cells by two longitudinal walls perpendicular to one another (figs. 58 and 63). By periclinal walls a central group of four cells becomes separated from four outer ones (fig. 64). The central cells divide no further. The outer ones divide by periclinal walls to form two concentric rows (fig. 65). The cells of both rows become eight in number, by anticlinal walls, the inner row undergoing no further division, but the outer, by another set of anticlinals, finally has sixteen. A fourth concentric row is then formed by periclinal walls from the outermost sixteen cells. The three inner layers consist of four, eight, and sixteen cells respectively, which numbers remain constant, but the fourth layer undergoes repeated divisions till a large number of cells are produced (fig. 67). These last lengthen greatly and form the wing of the scale. The surface view of the mature scale is seen in fig. 68, the longitudinal section in fig. 70. All of the cells but the stalk cells and the original basal cells undergo thickening of their walls in certain portions and lose their cell contents.
SCHIMPER (1) was the first to call attention to the water absorptive function of the scales, and his experiments along this line were so complete as to leave little else to be done. That the leaves of Tillandsia can absorb water is easily demonstrated either by wetting them with water and then watching it disappear, or by noting the weight before and after allowing them to remain a short time in water. That the channel of absorption is through the scales is shown by using colored water, which stains the stalk cells. Unlike most similar appendages of the epidermis, the scales do not hinder the leaf from becoming wet, but actually conduct water into the interstices beneath them. When dry, the leaf is of a gray color, due to the air enclosed by the scales, but when wet, the air is replaced by water, and a deep green color results. From an examination of fig. 70 it will be seen that the outer walls of the scale are thickened. When water is absorbed by the cells with thickened walls, they become turgid, expand below, and raise the wing of the scale well above the epidermis (fig. 69). The water absorbed by the outer cells of the scale passes to the stalk cells, which have thin walls and rich protoplasmic contents. Through these it passes through the basal cell to the waterstorage cells of the parenchyma. If the plant be soaked in dilute potassium iodide solution for a day, the walls of the stalk, basal, and neighboring parenchyma cells will be stained. It should be noticed that no ordinary type of epidermal cell with its thickened cuticularized wall separates the scale from the parenchyma. The cell that represents the epidermis beneath the scale is the basal cell resulting from the first division of the epidermal cell that gave rise to the scale. The walls of this basal cell are thin and uncuticularized. If a scale whose wing is raised well above the epidermis by the turgescence of its cells be treated with glycerin, the contraction due to loss of turgescence will draw the scale close down against the epidermis. This illustrates the process that takes place when scales become dry from evaporation, as occurs in nature.
Such a process cannot but assist the epidermis in checking transpiration, so that the scales may be considered not only as organs of absorption, but as serving to prevent too rapid escape of the water they have been instrumental in bringing into the plant.
The effect of an absorptive system extending over the entire surface has already been mentioned in the reduction of the mechanical and conductive tissues. As such reduction is found mostly in submerged hydrophytes, it will be seen that T. usneoides behaves in these respects much like such plants.
The scales stand in connection with the water-storage tissue. The cells of this tissue lie well distributed among the chlorophyll-bearing cells and keep them in a state of turgescence. Even after a plant has lost one-fourth of its weight by transpiration, and the leaves have become grooved by contraction, the chlorophyll-bearing parenchyma is unhurt. It is believed that the leaf shrinkage is due to a partial collapse of the storage tissue upon loss of water, rather than by decrease in turgescence of the green parenchyma. There is no evidence that the plant undergoes desiccation and subsequent revival, as in the case of Polypodium vulgare.¹
THE STOMATA.
In addition to protection afforded by scales, hairs, and thickwalled epidermal cells, xerophytes sometimes guard against too rapid transpiration by means of the position and structure of the stomata. Sunken stomata, or those vestibuled by an epidermal air space, itself with a narrow opening to the exterior, are all well known. In some xerophytic plants the usual closing of the pore by the guard cells is assisted in its function of checking transpiration by modifications in neighboring parenchymatous or epidermal cells. In Kingia australis, for instance, there is, according to TSCHIRCH,² a large intercellular space adjoining the stoma, partially filled with coiled cellular outgrowths of the parenchyma.
________________________________________________________________________
¹Since this paper went to press, one by MEZ (9) has appeared on the physiology of water absorption in certain species of Tillandsia, among them T. usneoides. MEZ corrects SCHIMPER'S observations as to the details of the absorptive process, claiming that the empty cells of the scale do not contain air, but are collapsed when the surface of the plant is dry. The thickened part of the scale swells when wet, raising it and causing the lumen to reappear in the collapsed cells. Water passes from exterior capillary spaces into the partial vacuum through thin places in the cell walls, whence, from the filled cells as reservoirs, the water is taken up and passed into the mesophyll by the stalk cells (Aufnahmezellen) through the usual process of osmosis. MEZ describes the scale of T. usneoides as having only one stalk cell instead of three. While it is true that two of the cells are very thin, their presence can readily be made out in good sections of mature scales and still more readily in sections of young ones.
² HABERLANDT, G., Physiologische Pflanzenanatomie. 2d ed. p. 399. 1896
The outgrowths do not stop, but merely hinder transpiration. Xanthorrhoea hastilis exhibits a similar contrivance. Camellia japonica and Prunus Laurocerasus have the faculty of filling up the airspace as a result of excessive drought or by death of the guard cells. In such cases tylose-like processes occur which block up all gas interchange. Pilea elegans differs from those meritioned above in that certain subjacent parenchyma cells develop thickenings on their exterior walls. One of these finally pushes up against the pore of the stoma and effectually closes it. There is no movement of the parenchyma cell away from the stoma, hence the aperture is permanently closed. From an examination of figs. 72 and 73 it will be apparent that Tillandsia presents a condition of affairs not widely different from that of Pilea. The principal difference lies in the fact that in Tillandsia the parenchyma cells undergo no thickening. Both longitudinal and cross sections through the leaf show outgrowths from the parenchyma cells lining the sides of the air space. The outgrowths turn upward and either stop up the opening of the stoma or else press directly against the guard cells. It will be seen that the enormously thickened walls of the guard cells preclude a possibility of change in their form. To show this experimentally some plants were placed in water and exposed to direct sunlight for a few hours. The leaves were then sectioned and the guard cells watched with a micrometer while glycerin was run under the cover glass. There was no measurable change. According to MEZ (6) the guard cells have lost the power of functioning, this power having been transferred to certain cells of the subjacent tissue which operate the passive guard cells, thus opening and closing the stoma. There are two cells which come in contact with the guard cell and may therefore be the means of moving it. One is the cell to which it is attached and which extends from the hinge to the inner face of the guard cell. This cell is usually continuous, but may be divided
by a cross wall into two cells. Should this cell, which is epidermal, become turgescent, it would tend to raise the guard cell, swinging its free side outwards. Such a movement, however, would close rather than open the pore of the stoma. The hinge is quite thick and may be much thicker than any shown in the figures. If the epidermal cell is divided the division wall would effectually hinder any movement of the guard cell. From these two considerations it would appear doubtful whether the guard cells move at all in either direction. Of course the glycerin experiment was repeatedly tried, but no motion was discernible. The only other cells which by contact with the guard cells can move them are the parenchyma cells whose processes push against the guard cells on the under side. It was at first thought that the parenchyma cells were operated by variations in turgescence of the epidermal cell, so that regarding the guard cells as immovable the epidermal cell would press downward upon the subjacent parenchyma cell during turgescence, and lower the process, thus unstopping the stoma. Out of a number of such processes only one reaches the center of the stoma, all the others being considered attempts that from necessity have failed. This explanation of the function of the parenchymatous outgrowth is plausible, to say the least, but it has not been experimentally proven by the glycerin test. Numerous instances were investigated carefully, but in not a single case did any of the processes change their position. It is here confessed that no reaction was noticed in any part of the stoma or adjacent tissue in response to the action of glycerin, nor was an instance found in fresh material where, the guard cells appeared to be separated. The experimental demonstration of the presence of a mechanism in the stomata, therefore, has not thus far met with success.
Another explanation might be mentioned, in which the processes are to be considered attempts on the part of the plant to close the stomata permanently. It may be that not all the processes actually reach the center of the stoma and close it, so that, granted that a small opening exists between the guard cells, the number of functional stomata would merely be reduced. The total number of stomata per square millimeter was ascertained and found to be relatively small. The estimate was made by counting the number of stomata in each section of serial sections taken from a portion of leaf of known length. For instance, a piece of leaf 3mm long contained 52 stomata. Calculating the surface from the circumference of the cross section, there would be 7 per square millimeter, or, in round numbers, 4,300 per square inch.-
It must of course be taken into consideration that sections of living leaves were used for experiment and not entire ones. If variations in the pressure of the water-storage tissue exert any influence on the opening and closing of the stomata it is very probable that the injury done to the tissue in sectioning would greatly interfere with the action of the mechanism.
HABERLANDT (8) figures the stoma of Tillandsia zonata,(Now Cryptanthus zonatus) which in respect to guard cells, and their supporting cells, resembles that of T. usneoides. The guard cells have greatly thickened walls, and a thickened hinge. From Haberlandt's account it is evident that he does not fully comprehend the mechanism. In T. zonata no subjacent parenchyma is mentioned as taking part in the opening or closing of the stoma.
THE STEM.
Aside from the vascular region, the stem differs in no essential particulars from the leaves as to structure. The stem, of course, has the added function of support, so that there is developed between and around the bundles a thick tissue of sclerenchyma fibers (fig. 74). The fibers measure about 750µ in length. They do not impart rigidity, but flexibility and power to resist longitudinal strain. If a fragment of moss is blown from one limb of a tree to another, and succeeds in getting a hold, the cortex of that portion of the stem that passes over the limb dies, and then disintegrates, leaving the sclerenchymatous axis, which holds the plant in place for several and perhaps many years. It is upon the durability and elasticity of this tissue that the economic value of the moss in upholstery depends.
What has already been said in regard to reduction in the function of the xylem and phloem of the leaves could with equal truth be said about the stems. With a superficial absorptive system and no root, the xylem as a conductive system is useless. The pendent habit and method of dissemination are both closely associated with reduction in mechanical tissue, though they are more likely to be the result than the cause of the reduction. The parenchymatous cortex, as in leaves, is supplied with chlorophyll-bearing cells, all of which are exposed to light, so that a tissue like the phloem, to carry elaborated materials to cells distant from the center of photosynthesis, would be unnecessary.
LOUISIANA STATE UNIVERSITY, Baton Rouge, La.
LITERATURE CITED.
1. SCHIMPER, A. F. W., Ueber Bau und Lebensweise der Epiphyten Westindiens. Bot. Centralbl. 17: 192 et seq. 1884.
--, Botanische Mittheilungen aus den Tropen. 2. Die epiphytische Vegetation Amerikas. pp.
162. pls. 6. Jena. 1888.
2. PEIRCE, G. J., On the mode of dissemination and on the reticulations of Ramalina
reticulata. Bot. Gazette 25: 404-4I7. 1898.
3. SMITH, R. W., A contribution to the life history of the Pontederiaceae. Bot. Gazette 25:
324-337. 1898.
4. HOFMEISTER, W., Neue Beobachtungen uber Embryobildung der Phanerogamen. Jahrb. Wiss. Bot. 1: 82-188. pls. 7-10. 1858.
5. MEEHAN, T., The Florida moss; Tillandsia usneoides. Proc. Acad. Nat. Sci.
Philadelphia 1875: 466.
6. MEZ, C., Monographiae Phanerogamarum. Editore et pro parte auctore Casimiro de
Candolle. IX. Bromeliaceae. Paris. 1896.
7. CAMPBELL, D. H., Studies on the flower and embryo of Sparganium. Proc. California
Acad. Sci. III. Bot. 1: 293-328. pls. 46-48. 1899.
8. HABERLANDT, G., Zur Kenntniss des Spaltoffnungapparatus. Flora 70: 97-110. p1. 2.
1887.
9. MEZ, CARL, Physiologische Bromeliaceen-Studien. I. Die Wasser-Oekonomie der
extrematmosphaerischen Tillandsien. Jahrb. Wiss. Bot. 40: 157-229. 1904.
EXPLANATION OF PLATES VIII-XI.
FIG. 2. Ovule fundament showing archesporial cell.
FIG. 3. Young ovule at period just before first division of archesporial cell.
FIG. 4. Spindle of first division.
FIGS. 5-6. Stages in formation of axial row of potential megaspores.
FIG. 7. Megaspores without separating walls. .
FIGS. 8-9. Enlargement of basal megaspore to form embryo sac mother cell.
FIGS. 10-14. Stages in formation of embryo sac.
FIG. 15. Pollen tube just after entering embryo sac.
FIG. 16. Fusion of polars before rupture of pollen tube: s, synergid; e, egg.
FIG. 17 Lateral discharge of pollen tube: e, egg; t, tube nucleus; s, synergids.
FIG. 18. Simultaneous double fertilization.
FIG. 19. Double fertilization with discharge of tube nucleus (t); e, egg.
FIG. 20. Fusion of male and endosperm nuclei.
FIG. 21. Ovule at time of completed embryo sac.
FIG. 22. Elongation of ovule and outer integument after fertilization.
FIG. 23. First division in formation of chalazal endosperm tissue.
FIG. 24. Chalazal endosperm tissue and portion of endosperm that is to serve as reserve
material in ripe seed.
FIGS.25-26. Two- and three-celled embryos.
FIG. 27. Formation of quadrant.
FIG. 28. Division of middle before terminal segment.
FIG. 29. Unusually early development of basal and middle segments.
FIG. 30. An unusual form of embryo.
FIGS. 31-36. Stages in embryo development; in fig. 34, the transverse walls in the terminal segment are oblique; the last three figures show beginning of dermatogen.
FIG. 37. Embryo about one-fourth grown.
FIGS. 38-40. Outlines of embryos in late stages of development; fig. 30 represents a mature embryo.
FIG. 41. Region in vicinity of growing point of a nearly ripe embryo.
FIG. 42. Root region of nearly mature embryo, showing dead cortical cells.
FIG. 43. Ripe seed.
FIG. 44. Barbs on hair of coma.
FIG. 45. Early stage in germination; outline of longitudinal section.
FIGS. 46-48. Stages in development of seedling; outline of longitudinal section.
FIG. 49. Longitudinal section through the growing point regions of a mature plant: s, sheath; st, stem; l, leaf; sa, stem apex; la, leaf apex.
FIG. 49a. Very young sheath.
FIG. 50. Cross section of leaf; p, pit in water-storage cell.
FIG. 51. Bundle of leaf enlarged to show phloem (p) and xylem (x).
FIG. 52. Megachloroplasts showing division into microchloroplasts.
FIG. 53. Stage in separation of microchloroplasts by which they become distributed through the cytoplasm.
FIGS. 54-61. Stages in development of the scale seen in longitudinal section; fig. 54 shows the epidermal cell from which the scale arises.
FIGS. 62-68. Stages in scale development seen from the surface; fig. 68 shows a mature scale.
FIG. 69. Scale in longitudinal section, after soaking in water for several hours; the wing is seen to be raised considerably above the epidermis.
FIG. 70. Scale in longitudinal section, drawn from a paraffin section; it will be seen to lie much closer to the epidermis than the one in fig. 69,
FIG. 71. General appearance of the surface of the leaf, showing the scales.
FIG. 72. Section through a stoma; the guard cells are unquestionably closed; in addition a process has grown up from the parenchyma into the pore of the stoma; s, scales.
FIG. 73. Section of stoma showing slight variation from that in fig. 72; figs. 72 and 73 were drawn from sections through living material.
FIG. 74. Cross section through the vascular region of the stem: p, phloem; x, xylem. —See Bot. Gaz.
- Culture and use: This species is very popular to vary the collection, because of its usnea-like habit, and is easily grown at a bright location in window or greenhouse; it is used as fill-up in packing.
Vernacular names : Guyana: Mora-hair, Saka-beard, Oldmans-beard, Spanish moss; Surinam: Jodenbaard. —See Gouda 1987
