Seeds of science

In a previous post, I shared some of my experiences hiking on Barro Colorado Island (BCI), Panama. I also delved into some thoughts regarding the drivers of plant biodiversity in lowland tropical forests. I discussed specifically Janzen-Connell effects, which postulate that natural enemies maintain tropical plant diversity by inhibiting plant growth near conspecific trees. Today, I would like to share a fun science project that two of my classmates and myself undertook on BCI, with the aiming of probing the forest for these Janzen-Connell effects. With 1 and ½ days for project completion, we found ourselves hiking to the far end of BCI, carrying almost 20kg of soil back to the lab, counting a thousand seeds, crushing rocks, dissecting plant embryos, and playing with fancy dies. And if you take the time to read this entire post, you might actually find that we have some cool results to share!

Janzen-Connell effects predict that tree establishment and growth should be inhibited near conspecifics. However, there remains much uncertainty with regards to which stages of the plant life-cycle should be most affected by this phenomenon. Is it as a sapling, or young adult, that trees experience challenges growing near other individuals of their own species? Or could it be on seedlings, just recently emerged from seeds, that Janzen-Connell operates? Or could it be even earlier, when trees are only embryos embedded within a seed?

Heather Stewart, Karthik Yarlagadda, and myself were curious about this. We wondered in particular if it could be that trees experience challenges to life near conspecifics as early as the seed stage, before any germination has even occurred. Seeds can indeed be vulnerable to many ailments, including direct predation by insects (such as weevils) or infections by soil fungi. For our class project, we therefore decided to ask: how does the density of seeds and their ability to produce seedlings change with distance from a parent tree? We predicted that seed density would decrease with distance from conspecifics, due to dispersal limitations. We also expected, as predicted by Janzen (1970), that seeds’ ability to mature would increase with distance from the parent. We received for this project the invaluable help and supervision of Camilo Zalamea, a STRI postdoctoral fellow specifically interested in the defenses of tropical rainforest seeds.

Our course instructors gave us 24hrs to perform all data collection: from Wednesday, 1pm, to Thursday, 1pm. With this time-limitation in mind, we first sought to identify the best study species for our project. Because we were interested in assessing seed density and viability, we decided that the ideal study species would: 1) be a pioneer species with a long-lasting seed-bank, 2) feature several easily-accessible and seed-producing adults on BCI, and 3) have seeds that are sufficiently large for easy collection and processing.

Finding a specific tree species that fulfilled all these criteria would not have been possible without Camilo’s in-depth knowledge of BCI trees and their seeds’ characteristics. The first species that we considered was Cecropia insignis, a pioneer and gap-dependent species (Brokaw 1986). However, Cecropia insignis seeds are extremely small and we were concerned that seed collection and processing would be too challenging within our limited time-frame. We then considered selecting Apeiba membranacea, another pioneer species with much larger seeds. However, Camilo mentioned that Apeiba seeds decompose as soon as they are no longer viable, and was therefore not confident in our ability to find non-viable seeds in Apeiba’s seed-bank. Finally, Camilo suggested that we work with Zanthoxylum ekmanii, a deciduous canopy tree that can grow up to 30m tall. Zanthoxylum ekmanii seeds are extremely resistant and look like little rocks. In fact, Dalling & Brown (2009) showed that Zanthoxylum seeds could remain viable in the soil for up to 18 years.


The bark (A), branches (B), seedlings (C), and seeds (D) of Zanthoxylum ekmanii, in the Rutaceae family. Photo credits: Heather Stewart (A and C) and Steven Paton (B and D).

We set out on a hike to the end of BCI’s Zetek trail in order to look for 3 mature Zanthoxylum ekmanii individuals that Camilo knew of. Forty-five minutes later, we found out with much disappointment that two of the three trees were no longer alive. One had likely fallen when Hurricane Otto hit Panama last November. Luckily, the recent nature of the fall made sampling for seeds possible within the framework of our class project. We therefore sampled seeds around both one live and one recently fallen Zanthoxylum.

At each tree, soil was collected along 30m-long transects starting at the base of the trunk. Soil was collected at 5m intervals (i.e. in 7 locations per transect) by digging out a cube of side 10cm. The top 2cm of soil sampled constituted “shallow” samples, and the subsequent 8cm were labeled as “deep” samples. Together, all 28 samples represented over 17kg of soil that we carried back to the BCI laboratory. Somehow Camilo was able to carry most of the soil and still leave us trailing behind him.


One of our sampling transects, starting at the base of our focal Zanthoxylum ekmanii tree and stretching 30m into the forest. Photo credits: Heather Stewart.

We spent the rest of our day weighing the soil samples, sieving them to remove excess soil, manually picking all Zanthoxylum ekmanii seeds from the remaining materials, and cutting all seeds in half. Believe it or not, we processed over 1,000 seeds this way.


Processing our samples. From left to right, sieving, extracting the seeds, cutting them in half, placing them in Tetrazolium chloride dye, and tallying test results using the microscope. Photo credits: Heather Stewart.


Viable (left) and non-viable (right) Zanthoxylum ekmanii seeds. Photo credits: Heather Stewart.

Once seeds were open, we were able to assess seed viability. This could be done visually: viable seeds were a shiny white on the inside, whereas non-viable seeds were brown, black, or simply empty. Seed viability could also be specified with higher confidence using a Tetrazolium chloride test. Tetrazolium chloride turns red within 24-36hrs if the seeds respirate, signifying viability. We therefore left our seeds in Tetrazolium chloride overnight in the hope that color changes would have started to occur by the following day.


Viable seeds colored by Tetrazolium chloride. Photo credits: Heather Stewart.

At noon on Thursday, we started scoring Tetrazolium chloride test results. Because we were not able to wait as long as is usually prescribed, none of the seeds displayed the bright red coloring that we had hoped to see. Nonetheless, a few had red hues already and harbored some beautiful pink colors. Exciting!

Now to some results! After entering all our data and producing some plots, we were able to draw a few conclusions from our work. Firstly, we were thrilled to see that our expectations with regards to seed density patterns were verified. Seed density peaked near the parent tree, before decreasing with distance from the seed source. This was not only exciting because it was consistent with Janzen (1970)’s predictions as well as previous studies performed on the same species (Dalling et al. 1997). It also meant that we had sampled female, and therefore seed-producing, trees – which we had not known for sure until then. It also comforted us that we had actually captured the seed shadow of our focal trees, alive or not!

We did not find such clear patterns for seed viability trends, likely due to small sample sizes for soils collected far from the parent tree (10m or more). More sampling would therefore need to be done in order to verify or nullify the existence of Janzen-Connell effects in Zanthoxylum ekmanii seeds. However, our data nonetheless hints at some potential patterns. When considering large sample sizes only (<10m from the parent tree), we see a moderate increase in seed viability with distance from the parent. Could this be due to higher densities of Zanthoxylum enemies in the vicinity of the parent? Only further exploration will tell. (I guess we will have to come back.)


Results from our project. Seed density (left) and percentage of seeds viable (right) are presented as a function of distance from the parent tree. It appears that seed density peaks between 0 and 10m from the parent tree, before decreasing exponentially with distance. No clear patterns in seed viability were apparent.

What are these “natural enemies” that could be affecting Zanthoxylum ekmanii seed viability near the parent, you might ask? Despite their thick and resistant coat, we saw many Zanthoxylum seeds featuring holes from tiny insect predators, such as weevils. Other seeds did not have any obvious cause of death, but could have been infected by a variety of soil pathogens and fungi. Alternatively, our seeds could simply have been rotting from excess soil moisture. This latter option would have nothing to do with Janzen-Connell effect as soil moisture is not specific to particular plant hosts. In order to perform a complete test of the Janzen-Connell hypothesis, this last possibility will therefore have to be ruled out. To do so, it will be essential to identify the cause of seed non-viability and to show that it is host-associated. This would entail learning to identify various types of seed damage under the microscope and cultivating fungi. More fun science to be done!

Spending time on Barro Colorado Island was once again a source of immense scientific inspiration. I would like to thank Camilo Zalamea, Erin Spear, Brian Sedio, and Scott Mangan for all the time they spent on the island with us, for all their insights, and for their mentorship. As always, I leave the island with a host of new questions in mind… and with a deep conviction that I must come back.


Leaving Barro Colorado Island on the Jacana boat. Photo credits: Chloé Debyser.



Brokaw, N. V. 1986. Seed dispersal, gap colonization, and the case of Cecropia insignis. Pages 323-331  Frugivores and seed dispersal. Springer. URL

Dalling, J. W., and T. A. Brown. 2009. Long-term persistence of pioneer species in tropical rain forest soil seed banks. The American Naturalist 173:531-535. URL

Dalling, J., M. Swaine, and N. C. Garwood. 1997. Soil seed bank community dynamics in seasonally moist lowland tropical forest, Panama. Journal of Tropical Ecology 13:659-680. URL

Janzen, D. H. 1970. Herbivores and the Number of Tree Species in Tropical Forests. The American Naturalist 104:501-528. URL


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s