A workflow for pollen identification.


The reproductive ecology of cactus is not well-studied. A small, side project of mine is to determine the pollinator guild of buckhorn cholla at Sunset Cove, Mojave Desert, and with which plant species, if any, it shares pollinators. The genera Opuntia and Cylindropuntia are known to be insect-pollinated, but I am curious which of the more than 659 species of bees in the Mojave Desert desert are pollinators.

As visitation does not necessarily lead to pollination, I removed the pollen loads from 22 bee visitors I caught during insitu observation periods. I also removed stigma from the cholla to quantify heterospecific pollen deposition i.e. evidence of pollinator sharing. Pollen ID is not easy task and so I have developed a workflow to make it more streamlined.


Prep a reference collection:

  1. Create a reference collection by removing pollen from the anthers of several flowers of every species blooming in the area. Store in ethanol.
  2. Mount and stain the pollen with fushcin jelly.
  3. Image each species of pollen grain at 3 magnifications. Measure the length and width of about 10 grains per species. I calibrated Lumenera’s Infinity Analyze software using a stage micrometer to make this really quick.
  4. Make a reference document to consult. I use a word doc where every page is a species. Add in the photos at several magnifications, the mean size and any notes.
Sample reference page for Echinocereus engelmanni (Hedgehog cactus)


To go through the stigma or bee pollen load samples, I use my Canon EOS 60D dslr with a 60mm macro lens pointed confocally into a light microsite at 100x. I used the remote shooting utility from Canon to control the camera with my computer and display the view onto a second monitor.

Home example of confocal setup
  1. I designate each coverslip on the slide as a zone and do 8 transects through each, counting the grains. Each line in my spreadsheet is a transect, each column is a species. I use 5 columns for buckhorn so I never have to count very high.
  2. I don’t count damaged grains, or grains in air bubbles.
  3. Each slide gets its own folder. I take photos of each heterospecific grain with the file name as the zone + transect + species, which is simple using the photo utility. Knowing where the grain is on the slide and what its surroundings are will be helpful if you need to find it again.
  4. The species can be tentative for now so don’t get too bogged down.
  5. Take photos of unknowns when first encountered and assign them morphospecies ID. I put these in a separate folder as a reference.
  6. Some species are easy to ID. Quite a few are not. The more grains you see the easier it is to spot the differences.
  7. To help ID, we can take a page from entomologists. Sort the photos by their tentative IDs, putting each species in a folder so they are visible all at once (do a bulk rename to append the folder name first). It is difficult to compare grains unless they are side by side, which isn’t realistic with one microscope.
  8. Sort until each folder contains identical grains, then assign them a species from the reference collection. Or assign them to a species group for species that are virtually identical (likely Asteraceae!). Assign any remaining to morphospecies. Update the datasheet with the corrections.  
Buckhorn cholla (larger) and silver cholla (smaller). Thankfully the most abundant grains are simple to differentiate.

Quick notes on plant harvesting

Ok, our competition trials never look that good.

useful paper: Designs for greenhouse studies of interactions between plants

Protocols

  1. Aboveground harvest: clip from soil surface.
  2. Belowground harvest: need to harvest entire plant at once by removing plants and roots from pot experiments (for instance) and gently washing to remove all soil but keeping roots and shoots intact.  Then, snip aboveground growth from below.
  3. Depending on level of replication and lowest possible independent sample unit, harvest one individual, one species, or all individuals of one species per pot into independent paper bags.
  4. Place in ovens at 68F for at least 2 days.
  5. Leave all plants in paper bags in oven until the moment you are ready to weigh.
  6. Remove from paper bags to weigh for small plants. Typically, I weigh to 3-4 decimal grams for small desert annual plants.
  7. Return plants to bag, do not return to oven, store in a paper box for a few weeks or until all data entered and checked.

#density series idea notes

Concepts

Do native species density trials at two scale – micro and mesoscale using pots and large plastic buckets. Mix of potting soil and sand best – 50:50.

Consider intra and interspecific competition series – with replacement, i.e. keep net densities per pot consistent.  Finally, perhaps consider competition with an exotic such as red brome.

Designs
with and without shade mimic in greenhouse
umbrella over buckets in mesocosm experiment
natives against brome in density series

Natives

a

b

c

then a*b, a*c, b*c

Or run each out in solo then against red brome.

 

Round 1.
a vs rb
b vs rb
c vs rb
at 1,2, 5, 10, 25 seeds etc or a simple with-replacement density series.
THEN, ready for it…. red brome from cali and red brome from Israel
Round 2. 
a+b vs rb
etc.

 

Scales

 

 

#Carrizo National Monument research update 2018_2

We had a more ambitious set of goals this season.

Goals

  1. Habitat use frequency estimates. Tools: a. telemetry of blunt-nosed leopard lizard with a total at least 1200 relocations split between AM/PM with an estimate of shrub-open and behavior. b. cam traps at shrub-open on still mode.
  2. Behavior estimates. Tools: a. cam traps on video mode at a total of 100 hours recording time. b. direct observation (with recording too) by humans of lizards and grasshoppers at a total of 100hrs.
  3. Shrub-plant-animal interaction estimates. Tools: exclosures at two sites to exclude different taxa in shrub-open mesohabitats. a. cages. b. cams c. vacuums. d. sweeps.
  4. Temperature profile estimates. Tools: a. pulse of collars on lizards b. loggers at microhabitat scales.
  5. Census grasshoppers. Tools: stick, sweep, and vacuum. Also do direct observation to assess whether they are significant consumers.

Teams

1a. Mario and Steph. Goals 1,2,3,5.

1b. Malory and Nargol. Goals 1,2,5.

2. Emily and Kat. Goals 1 & 4.

Extensions

Deploy one set of cam traps on still mode at a total no-shrub zone.

Get a solid handle on behavior by verts and inverts in the context of paired interactions with plants (at micro-scale) and shrubs at mesoscale.

Need an assay of insect diversity.

Animal sampling contrast protocol

Telemetry-scat design

Predictions to test

Do relocations map onto where scat is found too?

Is there scat fidelity from day to day?

Does telemetry relocation or conversely scat presence within a likelihood MCP (polygon estimating 95% percent change animals present within area) correlate with one another?

ie – imagine a telemetry polygon and a scat one too – never been done!! and then we statistically overlay them.

0.png

How to test

1. do relocations this saturday (day #1), enter into usual relocation datasheet attached.  Do morning and afternoon.

Enter data that evening and do a baton handoff to the scat team.

2. On sunday (day #2), scat people ‘check’ all locations where there was certainly a lizard.  In each day 1 relocation, is there scat the next day!!  So, imagine day 1 there were 100 spots where lizards where spotted. On day 2, what proportion of these have scat deposited.

3. On sunday (day #2), team telemetry repeats process and finds another 100 spots or whatever they can where they see lizards.  Same process – enter and hand off to team scat.

4. On Monday (day 3#), same process – team scat check day 2 telemetry relocations for scat – so we are holding as best we can scat scent cones etc to 1-day old and team telemetry repeats and finds new spots for next sampling.

5. On tuesday (day #4), team scat checks team telemetry relocations from the day before (day #3).

DATA OUTCOME

A total of 4 days sampling with 3 statistical days to test for scat detections with a 1-day lag where lizards were spotted. SO POWERFUL.

NOW — -as you can imagine – there are also a few bonus opportunities here…. 🙂

A.  If dogs have time, check back more than 1-day lag – ie on day #4, dogs can check ALL previous days days 3,2,1 – this gets the second main prediction – ie site fidelity. BE amazing to know this.

B. If team telemetry has the people and the receivers, it can also go the other way – team telemetry looks for lizards where there was scat detected – anywhere – the following day – so we pass the baton back and forth.

C. Team scat if they have time daily, checks other sites to fill in the region more – ie the polygon idea – to see how well regional NOT just point sampling works.