The Sphagnum air-gun mechanism resurrected? Not with a closer look


(Author for correspondence: tel +44 (0) 20 7 882 5291; email

When a long-held hypothesis is overturned by new experimental data (Duckett et al., 2009), it is only appropriate that the new evidence be examined critically and the experiments be repeated. Thus it was with real excitement that we read the letter by Sundberg (this issue, pp. 887–889), the acknowledged expert on spore dispersal in Sphagnum (Sundberg, 2005), that challenges our repudiation of the air-gun mechanism of spore discharge. However, we take issue with Sundberg’s conclusions and, even though he presents results that differ from those in our article, we argue that his interpretation of this additional evidence is flawed. In the process we highlight a hitherto unchallenged major mistake in Nawaschin’s (1897) interpretation of capsule anatomy that led to his original reasoning for the air-gun hypothesis.

Through a series of precise and comprehensive studies, we demonstrated that differential shrinkage of the cells in the capsule wall of Sphagnum causes the rigid operculum to pop off. The abundant pseudostomata on the capsule epidermis play a significant role in controlling and facilitating water loss (Beerling & Franks, 2009). The long-held view that air compression, not tension, is responsible for the explosive discharge of spores was refuted. Sundberg repeated our experiments and confirmed spore discharge from both pricked and unpricked Sphagnum capsules. In our experiments, based on a much larger number of sporophytes than those observed by Sundberg, both pricked and intact capsules had comparable spore dispersal patterns. However, unlike our reported observations, Sundberg demonstrated that spores were discharged further from unpricked capsules; unpricked capsules produced long traces of spores (Fig. 1a in Sundberg, this issue) whilst in pricked ones the spore simply spilled out of the capsules (Fig. 1b in Sundberg). The most plausible reason for these differences in spore discharge, as defined by Sundberg, is that pricking through the capsule walls quickens the drying process (as noted by both Sundberg and ourselves) and at the same time damages cell walls, thus distorting and lessening the even build-up of tension. Consequently, pricked capsules would less frequently exhibit discharge of long traces of spores compared with intact capsules. Moreover, in mature Sphagnum capsules, the individual dry spores are extremely light and lie within a narrow confined space (see later). Therefore, even the slightest change in capsule volume as lids pop off is sufficient to discharge spores a significant distance without having to invoke an air pressure of five atmospheres as claimed by Nawaschin (1897) and cited by Sundberg (this issue).

In this issue, Sundberg further states that unpricked capsules discharged spores with audible snaps. However, no details of the sound recording equipment are noted. Are we to believe that each of the capsules was heard by a researcher when it popped? How were the differences in sounds determined?

We have also observed capsule lid detachment mechanisms, which have never been suggested to involve pressure build-up, in a range of mosses with conical lids under the same conditions as in our Sphagnum experiments. Some species show forcible discharge while others do not. In Funaria hygrometrica, Bryum capillare and Bartramia pomiformis, the lids do not simply fall off but actively separate from the capsules to distances of up to 1 cm, even though the closed peristomes prevent spore discharge at the same time. Longer lids in species like Dicranum scoparium and Tortula muralis tend to be held in place for a time by the peristome after dehiscence. In the eperistomate species Tortula (Pottia) truncata and Weissia brachycarpa, spore discharge accompanies lid popping. All the aforementioned species have stomata (Paton, 1957; Paton & Pearce, 1957; and S. Pressel & J.G. Duckett, unpublished observations) that, unlike in Sphagnum, open into extensive intercellular air spaces, so lid detachment simply cannot be associated with the build-up of air pressure. It should also be noted that the pepper-pot apparatus in the Polytrichales (Parihar, 1961; Watson, 1971) is a very effective spore-dispersal mechanism; even the slightest pressure on the epiphragm is sufficient to forcibly expel spores from between the peristome teeth. No compressed air is required.

The major concern with Sundberg’s arguments is that they are based on Nawaschin’s (1897) illustrations of the structure of mature capsules before drying out. As originally illustrated by Schimper (1858), in Nawaschin’s Fig. 1, and thereafter often repeated in standard texts (Parihar, 1961; Watson, 1971), the dome-shaped spore sac in Sphagnum forms in the apical hemisphere of mature spherical capsules. By our calculation the spore sac has a volume of < 20% of that of the entire sporophyte. In striking contrast, Nawaschin’s Fig. 2 of a cylindrical capsule shows the entire apical hemisphere, or 50% of the capsule volume, occupied by the spore sac. This figure is quoted by Sundberg, who also cites volumes of 35 and 15%, respectively, for the air cavity and for the surrounding capsule tissue. Although the spore sac can be clearly seen in the upper hemisphere as darker than the air cavity below, when fresh transparent capsules are held up to the light (Sundberg, this issue), it is clear that the spore sac does not occupy the entire upper hemisphere. For this to be true, all the columella tissue inside the dome would have to break down. We are not aware of any published anatomical evidence that supports this interpretation, and Schimper’s (1858) illustration of a sporophyte on the point of changing shape shows the spores still confined to a narrow strip in the apical hemisphere. Moreover, our hand-cut sections of mature sporophytes about to change shape reveal that the central region of the apical hemisphere remains filled with thin-walled columella cells. All these illustrations and those cited by Sundberg are from capsules before drying out.

When the capsule dries, liquid is replaced by air in the cells below the spore sac and in the apical hemisphere, while mucilage disappears from around the maturing spores. As a result of the drying process, 80% of the internal volume of the capsule is air-filled at the point when the capsule changes shape; this is very different from the much smaller air chamber confined to the lower hemisphere of the capsule in Nawaschin (1897). Thus, Sundberg’s measurements, from squashed capsules, of emitted air amounting to 50–70% of the cylindrical capsule volume, with air remaining, equates closely with our 80% figure. Following our reinterpretation of sporophyte anatomy at the time of capsule contraction and dehiscence that identifies an 80% internal air space system, we can see no evidence for the presence of compressed air; in fact Sundberg’s new data strengthen the case against the air-gun hypothesis.

We also take issue with Sundberg’s interpretation of spore discharge in pricked capsules. Sundberg obtained similar results in relation to spilling of spores when capsules were pricked into the air space in their lower half compared with when they were pricked in the upper half, which he took to be into the spore sac. This is not surprising since pricking to a depth of 1 mm would puncture the air space in the apical hemisphere. We also compared dehiscence following apical and basal pricking and found no differences in spore discharge (S. Pressel & J.G. Duckett, unpublished) and therefore did not include these results in our article.

In contrast to the copious air emerging from cylindrical capsules, Sundberg also found that when five spherical capsules were pricked under liquid, air only emerged when the needle was removed. Squeezing demonstrated that at least one of them contained additional air. This he interpreted as the result of no or very little air compression in spherical capsules. If we were to accept that air pressure builds up when Sphagnum capsules change shape, then there should be the same amount of gas in both spherical and cylindrical capsules. We also observed bubbles emerging from holes pricked in cylindrical capsules at the point of dehiscence and also illustrated (Fig. 2f in Duckett et al., 2009) that there is a considerable volume of air present in the spherical capsules well before dehiscence.

We have an equally plausible explanation for less air emission from spherical capsules that does not involve pressure differences. In cylindrical capsules, the desiccated walls are hard and rather brittle, and thus pricking usually makes a vent through which air escapes. By contrast, spherical capsules have softer flexible walls that make seals around penetrating needles and reduce air escape, as recorded by Sundberg. These different responses to pricking of desiccated and hydrated Sphagnum capsule walls can be readily demonstrated from herbarium specimens. On adding water, dehisced capsules regain their spherical shape within minutes, as can be observed in nature (Fig. 2c in Duckett et al., 2009). When pricked, the hydrated walls seal the opening, whereas needles usually split the walls of dried capsules. The fact that only some of Sundberg’s spherical capsules contained air also suggests that these might have been immature and had yet to develop their internal air spaces fully.

As with lid popping, we conducted pricking experiments for the emission of air bubbles on two other mosses, Funaria hygrometrica and Bartramia pomiformis, with capsules of roughly similar dimensions as those of Sphagnum and obtained comparable results: lots of air bubbles emerge from holes in brown capsules on the point of dehiscence but far fewer from younger, fully expanded capsules, which have only just begun to dry out. Following up with squashing of the capsules reveals that the air volume is much the same in both.

We intend to follow up on Sundberg’s conclusion that there is a need for a more thorough quantitative investigation of capsule air volumes and a detailed study of internal structural changes during capsule contraction. When Sphagnum sporophytes become available we will conduct the suggested work alongside cryo-scanning electron microscopy studies of fractured sporophytes. This technique was used to explore gas-filled spaces in roots (Canny & Huang, 1993), bryophyte leaves and sporophytes (Duckett et al., 2010; Pressel et al., 2010) and embolisms in vessels (Canny, 2001a,b), and will provide direct evidence for the extent of gas spaces during the sporophyte maturation process. Our third approach will be to observe capsule dehiscence at different atmospheric pressures; if compressed air is involved in lid popping in Sphagnum then increasing the external air pressure should slow down and eventually stop the process. Whatever these future studies might reveal, our anatomical observations to date show that mature capsules below the spore mass are full of air-filled dead cells, many with broken walls, and hence the free emission of air bubbles following pricking. In addition, our published images show cracks in the outer walls of dried-out capsules, not to mention the collapse and breakdown of the walls of abundant pseudostomata (Beerling & Franks, 2009; Duckett et al., 2009). Thus we find it impossible to envisage how capsule walls might develop an air-tight seal, and how an air pressure of five atmospheres could build up in a system that is continously losing water through pseudostomata. We know of no biological system that is air-tight and loses water at the same time. Indeed, Sundberg states that an unresolved, intriguing puzzle is how Sphagnum capsules manage to build up air pressure while continuously losing water. Therefore, overall, in Sundberg’s experiments and conclusions we cannot find any convincing evidence for recommissioning the air-gun mechanism.


Financial support from a Leverhulme Trust Early Career Fellowship is acknowledged by SP. We thank S. Sundberg for his critique of our study: without it Nawaschin’s misinterpretation of Sphagnum sporophyte anatomy would have remained uncorrected.