Flower Morphology of Pelargonium odoratissimum and Pelargonium ionidiflorum Using Scanning Electron Microscopy



by Samantha Rothman
Smith College
Northampton, MA


Abstract:
    Scented geraniums have long been valued for their ability to mimic the scents of other plants. These scents are exuded by the true leaves of the plants. This experiment focused on the flower structure of Pelargonium odoratissimum and P. ionidiflorum. These flowers are often over-looked because they lack fragrance and are small. The samples were collected and fixed using a methanol/ hex dehydration series. The samples were mounted on to stubs and sputter coated. The viewing of the specimens was performed under the SEM. Variations in reproductive organ structure and sepal projections were found. More samples were taken and dissected under the dissecting microscope. The data collected from this phase of the experiment was most beneficial when comparing basic structure and secretions which may have been removed in the fixative process. The flower morphology of these two Pelargoniums was very similar, with differences in morphology being more apparent when viewed under the SEM.



Introduction:

    Pelargonium ionidiflorum and P. odoratissimum are both commonly referred to as scented geraniums, belonging to the family Geraniaceae (Reppert, 1989). Scented geraniums are native to South Africa and were first introduced to English horticulture in 1632. There are about 250 species of Pelargonium from South Africa, but of those only about a dozen count as scented geraniums (McDonald, 1996). After their arrival in England many more varieties have been bred (Martin, 1987). The leaves contain fragrant oils which mimic the odors of other plants, such as pineapple, orange, and apple (P. odoratissimum). Hence, they were given the common name of scented geraniums. These plants, particularly the rose scented varieties, were used commercially to produce oils used in fragrances. During the Victorian period they enjoyed great popularity and were collected heavily. In the United States the popularity of scented geraniums dwindled in the 1970's due to the increasing demand for hardier geraniums to be used as bedding plants. This became a threating situation during the 1980's as the rise of bedding plant type geraniums dominated the market and thus growers stoped the propigation of the scented geranium varities. This may have resulted in the lost of valuable germplasm for scented geraniums with in the United States (Ellis, 1995). Currently, there has been a resurgence in the popularity of scented geraniums.
     Although prized for their uncanny ability to mimic other plant scents, some are collected for their flowers. Because the blossoms of most scented geraniums are considered insignificant, those which posses bright blooms are included in collections regardless of the quality of the plant's (leaf) fragrance. Once such variety is P. ionidiflorum. Like its name suggests, it has rather showy flowers for a scented geranium.
     Pelargonium ionidiflorum and P. odoratissimum grow best in soil that is slightly acidic with a pH of 6.0 to 6.5. They enjoy full sun and even moisture, leaning toward the drier side. When cultivated indoors, insects may be a problem, such as white flies. This can be controlled with mild insecticides (McDonald, 1996).
     Pelargonium ionidiflorum and P. odoratissimum are in the family Geraniaceae which are dicots. The flowers of both Pelargonium ionidiflorum and P. odoratissimum are identical in structure in that the flowers are irregular, looking quite similar to the flowers of the common violet. The flower parts, including the pistil-leaves of the stigma, are found in groups of five with the exception of the immature stamens, where the young filaments are found in groups numbering seven on the flowers of both species (Gray, 1868).

Materials and Methods:

    Pelargonium ionidiflorum and Pelargonium odoratissimum were grown in the Scented Geranium Collection in the Show House at the Smith College Botanical Garden. The flowers were removed from the plants using tweezers and fine scissors. They were kept isolated by species and were immediately inserted into containers of methanol. The flowers soaked in the methanol for one hour, with the methanol being replaced approximately every 20 minutes, for a total of three changes. The methanol was poured off and ethyl alcohol (dehydrated 200 proof) was added. After 15 minutes, one specimen of each species was removed from the its original container and submerged in a smaller glass vial of hexamethyldisilazane, again remaining isolated by species. The hexamethyldisilazane was slowly evaporated overnight.
    The flowers were then attached to metal stubs using carbon paint. Once affixed they were placed in the sputter coater for two and a half minutes to be coated with gold and palladium. The samples were then viewed in the scanning electron microscope.
     The samples were later examined under the dissecting microscope after proving difficult to examine in the scanning electron microscope. Petals and some sepals were removed from each specimen using tweezers and fine scissors. More carbon paint was added to the stubs and remaining flowers. The specimens were again sputter coated for two and a half minutes. The samples were then viewed in the scanning electron microscope and photographs were taken.
    One week after the initial collection of flower samples (for SEM preparation), additional samples were collected from the same plants. They were examined and dissected under the dissecting microscope at 2.6 magnification.

Results:

P. odoratissium: When Viewed Using the Dissecting Microscope

    P. odoratissium was dissected under the dissecting microscope at 2.6 magnification. The flower had five petals in an irregular formation. The petals were distinct and arranged with two being longer , more up-right, and closer together so that they appear to be connected and thus form one heart shaped petal. The lower three were distinct and folded downward away from the summit of the axis and were hypogynous. The petals were all white with the exception of the two, back prominent petals. These had distinct markings in bright purple on their inner surface. The markings originated midway on the petals in a semi-solid bar which ran horizontal to the petals. Below these markings was a solid white space with the next set of markings being linear and running vertical into the base of the axis. These markings were in the identical color as the above markings.
    The sepals were distinct and folded back and downward from the summit of the axis and are hypogynous. The sepals were light green in color until they reached the axis, at which they were purple. The sepals on all flowers sampled numbered five.
     Some flowers sampled had no anthers, while the pistil was fully matured and there was pollen on the flower but it was not attached to the stigma. The pollen grains found were cylindrical in shape and colored orange-red.
    All flowers sampled had seven filaments, regardless of their maturity. The filaments were solid white in color. The pistil on all flowers sampled had hair like projections in a thick mass around the outside of the base of the ovary. The stigma retractions numbered five. The tops of the retractions were bright magenta in color with a glistening, clear coating. The top two-thirds of the outer surface of the pistil was pink/ purple in color.

P. odoratissium: When Viewed Using SEM

    The stamens of P. odoratissium rise vertically covering the immature pistil. The filaments are distinct and stay close to the surface of the ovary. They branch outward after passing the area of the style. The filaments have a hypogynous insertion. The sepals surround the base of the pistil and stamens. The sepals are joined to the receptacle, which was used to hold the flower during the various fixative processes and therefor is damaged. The sepals once surrounded the flower when it was in its bud stage and have now folded back as the flower blossomed. They curl backward from the reproductive area without rolling up on themselves. This leaves the inner surface exposed upward with the outer surface now underneath. The base of the ovary is covered with coarse hairs, protected by the surrounding filaments (Fig. 1).



Fig. 1 - Pelargonium odoratissimum flower with petals removed. Accelerating voltage: 3.5 kv. Working distance: 25 mm. Probe Current: 13. Magnification: 17x . Bar represents 1 mm.


    The sepals of P. odoratissimum when viewed at 100x magnification showed the surface to be covered with hair like projections (Fig. 4a). These hair like projections were tapered in shape, with the connection to the sepal being the widest point and tapering to a fine tip at the end of the hair. The hairs were straight and smooth in texture, with a slight bend mid-way up. The hairs appeared to be spaced evenly, with approximately 150 microns between each, but in no apparent pattern or order (Fig. 2).
    When magnified to 150x, additional projections were found that were shorter in length with a solid primary base, connective type stem, and a bulbous top portion (Fig. 3). The primary base of the projection was approximately 26 microns in length and 26 microns in diameter across its middle. The connective type stem was approximately 15 microns in length and 14 microns in diameter. Furthermore, the connective type stem appeared to be composed of more dense material than the primary base. The top portion was textured with crisp ridges (Fig. 3). The top portion of the projection in Fig. 3 was standard in size when compared with other projections viewed. The bulbous portion of the projection with the crisp ridge texture was approximately 26 microns in diameter. Not all of these projections had the crisp ridged coat attached. Those whose outer coatings were unattached typically had the coating adjacent to it, on the surface of the sepal. The top portion which was then revealed was cylindrical in shape and completely smooth in texture.
    The cells of the epidermis of the sepal were round in shape with a vertical rise which gave them a plump appearance. They were textured, with their markings being similar to that of the human finger print (Fig. 3, lower right corner). When a sepal was removed from the sample and re-coated, the cell walls were exposed. The cells of the lower and upper epidermis were four sided with defined corners and arranged in a linear fashion. The cells of the mesophyll were generally larger and more oval in shape (Fig. 4 and Fig. 2).




Fig. 2 - P. odoratissimum sepal at 150x magnification. Accelerating voltage: 3.5 kv. Working distance: 8 mm. Probe Current: 11. Bar represents 150 microns.



Fig. 3 - P. odoratissimum sepal at 1,300x magnification. Accelerating voltage: 3.5 kv. Working distance: 8 mm. Probe Current: 11. Bar represents 13 microns.




Fig. 4 - P. odoratissimum sepal cell walls at 900x magnification. Accelerating voltage: 2 kv. Probe Current: 11. Working distance: 8 mm. Bar represents 36 microns.



Fig. 4a - P. odoratissimum sepal and view of cell walls at 100x magnification. Accelerating voltage 2 kv. Working distnace: 39 mm. Probe current: 10. Bar represents 200 microns.


    The cells on the filament are arranged vertically in a row -like formation with a ridged surface texture. The texture of the surface cells appears to be rather consistent. The ridges run parallel to one another in a vertical orientation. Furthermore, the cells appear to be more compact as the filament enters the back of the anther, to its location of attachment. The connection between the anther and the filament is in a cavity. The filament recesses into the anther to form the connection (Fig. 7).
    The anther is attached to the filament in the versatile position, although the connection is more secure than that of the versatile position commonly seen in flowers such as lilies (Fig.1). The anther turns inwards in the incumbent position. The back side of the anther, where it joins with the filament, has surface cells arranged in row like formation, running vertically and parallel to one another (Fig.7). They appear to lap over one another in a woven pattern. The cells appear to be fuller with more vertical rise than those of the filament. Their surface texture is ridged with the ridges initiating on the side of the cell and running horizontally across. At the entry point of the filament the cells are arranged in complete curved rows. As the rows become closer to the interior of the anther they become more linear and have greater definition with less of a woven pattern.
    The anther (Fig. 5) is dehiscent and has one pollen grain left attached. Not only is this apparent due to the presence of the pollen grain, but from the folded back position of the upper surface of the anther itself. The pollen grain is cylindrical with two apparent clefts on it surface where the net-like outer coat becomes less porous. The holes in the outer coat appear to be four sided (Fig. 6).





Fig. 5 - P. odoratissimum anther with pollen grain at 120x magnification. Accelerating voltage: 3.5 kv. Working distance: 8 mm. Probe Current: 11. Bar represents 120 microns.



Fig. 6 - P. odoratissimum pollen grain at 1,500x magnification. Accelerating voltage: 3.5 kv. Working distance: 8 mm. Probe Current: 14. Bar represents 15 microns.



Fig. 7 - P. odoratissimum filament connecting with the back of the anther at 300x magnification. Accelerating voltage: 2 kv. Working distance: 8 mm. Bar represents 300 microns.


P. ionidiflorum: When Viewed Using the Dissecting Microscope

    P. ionidiflorum was dissected under the dissecting microscope at 2.6 magnification. The flower had five petals in an irregular formation. The petals were distinct and arranged with two being longer , more up right and closer together so that they appeared to be connected and thus form one heart shaped petal. The lower three were distinct and fold downward away from the summit of the axis and were hypogynous. The petals were all solid colored in a bright pink/ purple with the exception of the two, back prominent petals which had distinct markings in dark purple on their inner surface. The markings originated midway on the petals in a semi-solid bar which ran vertical to the petals and continued into the axis.
    The sepals were distinct and folded back and downward from the summit of the axis and were hypogynous. The sepals were very dark purple in color, throughout. The sepals on all flowers sampled numbered five. Where the sepals joined with the axis, a clear liquid was present.
    All flowers sampled had no anthers on the filaments if the pistil was fully matured. On all flowers sampled seven filaments were found in various sizes. The filaments were light pink in color at their tops, while the color faded to white toward the axis.
     There was a small quantity of pollen on the flowers sampled, but it was not attached to the stigma. The pollen grains found were smaller than those of P. odoratissimum and less cylindrical in shape. The pollen found was orange colored .
    The pistil on all flowers sampled had hair like projections in a thick mass around the outside of the base of the ovary. The pistil was colored dark magenta, with the retractions being the darkest portion of all. The color was graduated slightly, with the lightest shade being found where the pistil joins the axis.The stigma retractions numbered five.


P. ionidiflorum When Viewed Using SEM

    The filaments of P. ionidiflorum were lacking in anthers and were shorter in length than the pistil. The filaments had a hypogynous insertion. Several filaments were removed to allow better viewing of the outer surface of the ovary. The remaining portions of the filaments can be seen in Fig. 8 and should not be confused as smaller filaments (Fig. 8).
     The pistil was the most prominent part of the reproductive structures. The pistil had five stigma retractions. The outside of the ovary was covered in hair-like projections which had a distinct pattern. They appeared to be laying on one another, growing upward in the direction of the stigma, forming a hair-like mat. The sepals surround the base of the pistil and stamens (Fig. 8).
    The sepals join the receptacle (which was used to hold the flower during the various fixative processes and thus has been damaged). The sepals surrounded the flower when it was in the bud stage and then folded back as the flower blossomed. The sepals curl back from the reproductive parts, without becoming shriveled and curling on themselves. The inner surface of the sepal is now on the upper side while what was once the outer coat of the bud is now underneath, facing downward. The inner surface of the sepal appears to be smooth in texture with no apparent projections (Fig. 8).




Fig. 8 - Pelargonium ionidiflorum flower with petals removed at 17x magnification. Accelerating voltage: 3.5 kv. Working distance: 25 mm. Probe Current: 13. Bar represents 1 mm.


    The stigma of the pistil had five retractions. There was an apparent center cavity which runs down the center of the pistil. The retractions were of a solid tissue which has a compact cellular structure running in a linear fashion. This tissue runs vertically, originating from the center of the cavity, and down the length of the retractions (Fig. 9). The tops of the retractions have distinct formations of large bulbs. The bulbs have a smooth surface texture. Some of the bulbs have openings in them with a substance coming out from with in the structure. These bulbs appear to be affixed to the stamen retractions by a stem like structure (Fig. 10).




Fig. 9 - P. ionidiflorum stigma at 220x magnification. Accelerating voltage: 3.5kv. Working distance: 37 mm. Probe Current: 11. Bar represents 220 microns.



Fig. 10 - P. ionidiflorum stigma at 850x magnification. Accelerating voltage: 3.5 kv. Working distance: 8 mm. Probe Current: 12. Bar represents 85 microns.



    The epidermis of the sepals has cells that run in a vertical fashion with twisted, hair like projections. These hairs are not arranged in any particular order. The hairs are not totally round and seem to have areas which are flat ( Fig. 11). When the sepals are under higher magnification the epidermal cells have a distinct texture which is similar to the human finger print, with lines running down the length of the cells. Embedded in spaces between the cells are pairs of guard cells, which form the stomata. These cells are smooth in texture with a single ridge running between the two guard cells. There is an opening in the center, between the two cells. The hair-like projection originates from in-between two epidermal cells. This hair has a flat section close to its point of origin. It then forms a loop and continues upward (Fig. 12).





Fig. 11 - P. ionidiflorum sepal at 100x magnification. Accelerating voltage: 2 kv. Working distance: 39 mm. Probe Current: 10. Bar represents 100 microns.



Fig. 12- P. ionidiflorum sepal at 250x magnification. Accelerating voltage: 3.5 kv. Working distance: 25 mm. Probe Current: 11. Bar represents 25 microns.



Discussion:

    The fixation of the samples for viewing under the scanning electron microscope proved to be adequate for this experiment. The method of fixation was chosen based on the work of Neinhuis and Edelmann (1996) who experimented with various forms of fixation of plant reproductive organs. They found that fixation using the methanol technique applied here worked best. One difference in fixation between that experiment and this was the use of the critical point drier. In further research, the use of the critical point drier was pervasive.
    Although the samples, after under going size reduction, were able to be viewed accurately, it is possible that using higher accelerating voltage could have increased detail and viewing capacity. While these samples were viewed using an accelerating voltage of 3.5 kv, flower samples such as apple blossoms have been viewed using accelerating voltages as high as 15 kv (Mansvelt and Hattingh, 1989). It could be possible that they were able to use the higher voltage because they made their samples smaller by dissecting the flower parts previous to the sputter coating. They also and sampled the reproductive parts individually (Mansvelt and Hattingh, 1989). When floral organs of Brassica napus were studied, higher accelerating voltages were used as well. Polowick and Sawhney (1986) were able to use voltages as high as 30 kv.
    The flowers of P. odoratissimum and P. ionidiflorum are identical in morphology when comparing the numbers of parts and their arrangement within the flower. The flowers of P. odoratissimum are smaller and less brilliant in color than those of P. ionidiflorum. Furthermore, the anthers were the most prominent reproductive structure on P. odoratissimum, while on P. ionidiflorum the stigma was the most prominaite. This could be due to the time at which the flowers were sampled and further studies would be necessary to determine if there was a more dominate reproductive organ with in the flowers.
    Through dissection of the flowers under the dissecting microscope and viewing the samples under the SEM it is apparent that the reproductive parts come into maturity at different times suggesting that the plant seeks to be cross pollinated with another plant rather than be in-bred.
    The anther of P. odoratissimum is dehiscencent showing that the male portion of the flower has come to its sexual maturity, while the pistil was not fully matured or had already passed its prime. This further supports the theory that both Pelagonium species exhibit dichogamy.
     The pollen grain of P. odoratissimum has a net exine structure (Iwanami et al, 1988). The intine structure was not able to be viewed using the SEM. The pollen had a more cylindrical shape than that of P. ionidiflorum and was larger in size.
    The stigma retractions of P. ionidiflorum had distinct structures on them which appeared to be secreting some type of substance. This could be some type of glue which helps the pollen grain adhere more readily. Another possibility is that it could be some type of nectar which is attractive to insects an therefor would aid in flower pollination . Furthermore, a type of clear substance was seen on the stigma of P. odoratissimum when viewed with the dissecting microscope. This could be the same substance being exuded by P. ionidiflorum. Because the flowers of both species appear to exhibit dichogamy it is highly possible that this substance is meant to attract some creature to aid in the cross pollination of the plants, since self pollination would be unlikely.
    The sepals of P. odoratissimum and P. ionidiflorum are note worthy because they are modified leaves and therefor can give some clues into the morphology of the true leaves. As both these plants are scented geraniums, they exude oils from their leaves giving off the scents that they are so famous for. On the sepals of P. odoratissimum a bulbous structure was found in addition to the hair-like projections. It is possible that this structure could be a form of an oil gland. During a study of the effects of dust on leaves, P. coccineus was observed as having gland structures on its leaves, further suggesting that the structures on the sepals of P. odoratissimum could possibly be glands (Eveling 1986). As the name ÒodoratissimumÓ suggests, the plant is one of the most fragrant of the scented geraniums. This is in contrast to the P. ionidiflorum, which is a scented geranium, but one of lesser fragrance. This could explain why there were gland structures on the sepals of one, but not on the other.
    The hairs found of the sepals of both plants were different in morphology, number, and arrangement. It is not possible to say if hair-like structures are present on all Pelargonium sepals, but they have been found on P. coccineus, as well (Eveling, 1986). Furthermore, sepals do not always have these structures. Brassica napus is an example of such a flower (Polowick and Sawhney, 1986).
    In addition to projection morphology, the stomatal arrangement was different between the two species. While the stomata of P. ionidiflorum were pronounced and easily seen, the stomata of P. odoratissimum were not viewed at all. This could be due to the untrained eye of the viewer or because they are significantly smaller. In studies done by Mansvelt and Hattingh the substomatal chamber was able to be viewed using an accelerating voltage of 15 kv. Perhaps if the sample of P. ionidiflorum had been smaller and a greater voltage used, the viewing of the substomatal chamber would have been possible.

Conclusion:

    Through out the research used for this experiment, one major difference in the methods of fixation between their experiments and this one was the use of the critical point drier. They might have employed this technique to insure the original texture of delicate structures such as petals. This study hope to view petal morphology as well. Unfortunately this was not possible due to the problem of ÒchargingÓ on the specimen and the petals blocked the view of the reproductive organs. The petals could have been viewed if the specimens had been dissected prior to being mounted on the stub, as was done with the apple flower specimens (Mansvelt and Hattingh, 1989). Further experimentation with fixation technique would be advantageous to establish a technique which would provide maximum results using the SEM.
    It is apparent that the flower morphology of both of species is highly similar. Their differences to the unaided eye are size and color. When dissected under the dissecting microscope their reproductive differences become more clear. The more dramatic differences, such as the morphology of the hair like structures found on the sepals, became apparent when using the SEM.
    The presence of the gland type structures on the sepals of P. odoratissimum and not on those of P. ionidiflorum could be directly related to the weaker leaf scent of P. ionidiflorum. Testing of the leaf structure of both of these plants would be helpful to confirm this.




Literature Cited

Ellis, David J. 1995. A passion for pelargoniums. American Horticulturist. 74:30-33.

Eveling, D. W. 1986. Scanning electron microscopy of damage by dust depositis to leaves and petals. Botanical Gazette. 147:159-165.

Gray, Asa. 1878. GrayÕs Lessons in Botany. Ivision, Blakeman, Taylor & Co.:New York.

Mansvelt,E. Lucienne and M. J. Hattingh. 1989. Scanning electron microscopy of &invasion of apple leaves and blossoms by Pseudomonas syringea pv. Syringae. Applied and Enviormental Microboilogy. 55:533-538.

Martin, Joy Logee. 1987. Scented Geraniums. Brooklyn Botanic Garden Record, Plants & Gardens. 43(1):56-57.

McDonald, Elvin. 1996. Pelargonium pleasure: scents & sensibility. Plants & Gardens.

Neinhuis, C. and H.G. Eldelmann. 1996. Methanol as a rapid fixative for the investigationg of plant surfaces by SEM. Journal of Microscopy. 184(1):14-16.

Polowick, P.L. and V.K. Sawhney. 1986. Scanning electron microscopic study on the initiation and development of floral organs of Brassica napus. American Journal of Botany. 73(2): 254-263.

Reppert, Bertha. 1989. Scented geraniums, mimic plants. Brooklyn Botanic Garden Record, Plants & Gardens. 45:72-77.

White, J.W., et al. 1990. Floral initiation and development in Aquilegia. HortScience. 25(3): 294-296.



Additional Information on Pelargoniums can be found at these pages:

Plant Link to Pelargonium
Miles Estate Herb and Berry Farm (This is a company I know nothing about, but it can give you an idea of the variety of scented geraniums that are available in the market place).

And for anyone who is concerned with the rapid loss of valuable medicinal plants in their native habitat, please check out this page.
United Plant Savers