In each of four regions (Chiapas, Guatemala, Honduras, Nicaragua) we will choose nine sampling sites: 3 low elevation (0-500m), 3 mid-elevation (1000-1500m), and 3 high elevation (2000m and above). We will travel from one site to the next, carrying out a replicated program of field sampling at each site.
The emphasis is on evergreen wet forest habitats. Study sites are selected based on availability of evergreen wet forest.
Sierra Morena (mid elevation): a small community in the biosphere reserve La Sepultura, at the north end of the Sierra Madre de Chiapas. The community itself was on an ecotone between dry deciduous forest and evergreen forest. The evergreen forest was a mosaic of different land use histories. Some areas were open coffee farms with sparse shade trees. Some areas had a closed canopy of secondgrowth trees with an understory of coffee. Some areas had oldgrowth forest canopy with an understory of either coffee or more often a palm (Chamaedorea) planted for the ornamental foliage trade. There were few areas of old-growth forest where both canopy and understory were intact.
Quetzal (mid and high elevation): Campamento Quetzal, a research and education facility near the small town of Custepec. It is in the biosphere reserve El Triunfo, at the southern end of the Sierra Madre de Chiapas. The land within a few kilometers of the building is a mosaic of old-growth forest and about 40yr old secondgrowth forest. Campamento Quetzal was used as both a mid-elevation and high-elevation site. The area around the building itself, from 1500-1700m, was used as a mid-elevation site. A ridge crest a few kilometers from the station was above 2000m and was used as a high elevation site.
Coapilla (high elevation): a small town in the central Chiapas highlands, north of Tuxtla Gutierrez. This is a highly fragmented landscape, mostly agricultural land and pine/oak forest, with only a few small patches of evergreen forest remaining. Sampling was carried out in two forest patches, between 1900 and 2000m elevation.
Huitepec (high elevation): Sampling was carried out at two sites near San Cristobal de las Casas in the central highlands. One site was Reserva Huitepec, a medium-size reserve with old-growth montane wet forest. Another site was Bazom, a forest fragment several kilometers south of San Cristobal. Both sites were about 2500m elevation.
Nah‡ (mid elevation): an indigenous reserve on the northern slope of the central highlands, at 900-1000m elevation. It is an area that is extensively forested, with a mosaic of different degrees of disturbance from selective logging and milpa formation.
Metzabok (low elevation): an indigenous reserve a short distance and downslope from Nah‡, with elevation 570m. There are medium-sized patches of old-growth forest around a lake. The surrounding landscape has mostly been cleared for agriculture.
Both Nah‡ and Metzabok are islands of forest that were formerly continuous with the much larger Lacandon rainforest to the south. They have become forest islands within the last 50 years.
Paso Naranjo (low elevation): an ecotourism establishment on the Rio Tulij‡, near Salto de Agua, in the northern lowlands of Chiapas. The landscape was a mosaic of forest patches and milpas (agricultural clearings). The forest patches were closed-canopy secondgrowth growing on very rocky land. The surface was highly irregular, with large limestone boulders, ledges, basins, and cracks. Elevation was 50-100m.
Play—n de la Gloria (low elevation): a small community on the southern edge of the Montes Azules biosphere reserve, on the Rio Lacantœn. Some sampling occurred across the river from the community, in an area of mostly old-growth tall rainforest. Some sampling occurred on the community side of the river, an area of mixed agricultural land, second growth forest of various ages, and small remnants of mature forest. Elevation was 150-200m.
To review collection gps points on Google Earth, open Google Earth, go to "Add" on menu bar, select Network Link, paste in one of following url's.
Terminology
Sample: a single miniWinkler or maxiWinkler sample. Quadrat: the 1m2 quadrats used for miniWinklers. Transect: 245m long transect, 50 miniWinkler samples at 5m spacing. Site: a locality where a pair of transects is located.
General Procedures
Day 0 (arrival). Get established in designated housing. Locate and establish site for hanging 100 miniWinkler and about 20 maxiWinkler samples. Establish two sifting teams of 5 people each. Prepare backpacksÊwith sifting kits: 2 sifters, 2 machetes, 50 sample bags, trowels, preprinted collection code labels and data sheets, clipboard, scissors, 50 Whirlpac bags, tape measure, compass.
Day 1. Each team goes to designated miniWinkler transect site.
MiniWinkler transect: select starting point and direction of transect. Orient transects to stay within forested habitat. One person holds free end of 50m tape at starting point; second person carries tape reel and sighting compass to play out 50m tape to end. Sampling carried out on up-tape (away from start point) right side of tape, so return to start point either along tape itself or on up-tape left side, to limit disturbance of quadrats prior to sifting. Move along tape, sifting quadrats at 5m intervals. To precisely and consistently locate quadrats, carry meter stick. At 5m interval, place meter stick perpendicular to tape on up-tape right side. This marks one corner of 1m2 quadrat. Quadrat can be sited by eye using meter stick.
Use machete to macerate litter, soil, dead wood in quadrat. Sift immediately after maceration, using trowel or gardening tool to scrape litter into sifter. Sift all quadrat litter if resulting siftate is less than 3 liters. If litter is deep and abundant, such that sifting produces more than 3 l, subsample different portions of quadrat until maximum volume is reached. Place siftate in cloth bag. Cut collection code from sheet, place in Whirlpac bag, and place Whirlpac bag in cloth bag with sample. It is important to cut labels out as samples are taken in the field, rather than cutting them ahead of time, to assure maintenance of sample order and sample number. Shake sifter vigorously to remove any loose litter, to minimize contamination among quadrats, and continue to next quadrat. Guard samples in shade along the transect for gathering on return. Be sure to always keep samples in shade. Avoid compaction of samples.
When end of 50m tape is reached, use compass to continue the transect for the next 50m, dragging tape reel another 50m.
If a quadrat is partially covered by vertical tree trunk that is greater than 30cm diameter at ground level, shift quadrat to edge of trunk.
For each quadrat take litter depth data by measuring approximate litter depth (in whole cm) at each of four corners of quadrat. Centimeters can be marked on machete blade, which makes a good ruler. Record habitat and canopy condition (e.g., closed canopy mature wet forest). Make other brief notes on quadrat characteristics (e.g., moss mat from canopy in sample, shifted due to large tree in quadrat, stones or rocky surface, rotten log).
Obtain precise GPS reading of transect start point. Record in detail any changes in direction of transect, so that precise spatial map of quadrats can be constructed.
If transect comes to edge of forested habitat or otherwise impassable barrier, transect can "reflect," turning at an arbitrarily selected angle and continuing in a new direction. Choose angle that is approximately 45¡ or greater, so that transect does not double back too closely on itself. Do not decide to just go along a trail because it is easier.
MaxiWinkler sample: sift additional litter in general vicinity of miniWinkler transect. Select pockets of litter and rotten wood, chop with machete, and sift. When about 10 l obtained, place siftate in cloth sack in backpack and add collection code. Continue sifting until five or more bags obtained, or as time permits. Record data on location and general habitat characteristics of sifting area.
One person erects two Malaise traps in general vicinity of transect. Precisely georeference the Malaise traps. Harvest sample and take down Malaise traps at latest possible time, on the last day prior to departure for next site.
Rain: do not sift until 4hr or more after the last heavy rain. Do not sift during rain. If it starts to rain during sifting, suspend operations and wait until 4hr after rain has stopped. If rain causes a delay in sifting, schedule should be adjusted to always allow the full 3-days of extraction. The extraction period should never be shortened.
Return to Winkler-hanging facility and immediately hang samples in Winkler bags. Attach dry cup to Winkler bag. Assemble Whirlpac bag with collection code onto plastic cup sleeve. Place siftate in bucket. Scoop litter into mesh sack always over bucket (there is a rapid exodus of arthropods immediately upon placement in mesh sacks, so in the initial hanging phase any time a mesh sack is not over a bucket or Winkler bag it is suffering significant loss of material). When mesh sack is nearly full, rapidly transfer from bucket to Winkler bag. Tap any remaining loose material remaining in bucket into open mouth of Winkler bag. Jiggle Winkler bag so that loose material drops into dry cup. Quickly switch dry cup with another one. Pour accumulated dry material into top of mesh sack. If loose material continues to fall into dry cup, repeat process multiple times. Close upper Winkler bag opening with cloth sleeve. Add 95% ethanol to Whirlpac, remove dry cup and hold lower mouth of Winkler bag closed while tapping contents of dry cup into Whirlpac. Attach cup sleeve with Whirlpac to Winkler bag. Avoid any bumping of Winkler bags from then on.
Carry out similar procedure for maxiWinkler samples. Load litter in maxiWinkler bags, 4 mesh sacks per bag. Record total number of mesh sacks filled for sample.
Day 2. Teams return to transect sites. Two people begin baiting protocol, two people begin beating protocol.
Baiting protocol: Select portion of path about 100 paces long. Use 3x5 cards cut in half as baiting platforms. Lay out 20 platforms along edge of trail, at 3-pace intervals. For each platform place bait on platform. Sprinkle a small amount of bait off edge of platform. Periodically check baits, cycling among them over a 2 hr period. If only one or two ants present, do not disturb. When larger recruitment occurs, use pooter to obtain sample into Sarstedt vial and leave vial near card.
Pheidole is a genus of special concern. Pheidole have a discrete major worker caste that is important for taxonomic work. Major workers are more timid and much less abundant than minor workers. A bait recruiting many minor workers may slowly accumulate a few major workers. When pooting, go for the major workers first. At the beginning of the recruitment process, check closely for major workers. If a major worker is seen, attempt to poot it while minimizing disturbance to other workers, so that they continue to recruit. Often there is a peak in worker abundance over time, followed by a diminution as the bait is removed. Judge when peak abundance occurs and at this time attempt a fuller sample of minor and major workers. Try to poot just workers, avoiding getting much of the bait in the vials. The oil in the bait tends to coat specimens, which is bad. At the end of the sampling time, look under the platform for any workers there.
At the end of the sampling period, place vials (only those with ants; for baits where no ants arrived, reserve vials for future use) in plastic bag containing label identifying collection code, date, time, location, habitat, and collector. Do not put bait cookie in same bag with vials. Gather bait cards when finished.
In the field, a single collection code will be assigned for each batch of vials from a single person's baiting session, of the form m-c-xx-y-zz, as described below. Later, during sample processing, a subnumber will be added to make a separate collection code for each vial.
The bait of choice is crumbled Pecan Sandies.
Beating protocol:ÊSelect a portion of a trail and start beating along the sides of the trail and nearby vegetation. ÊMaintain a collecting vial with ethanol and a collection code. Collect from head height to just above ground level on any dead or living vegetation present.Ê Beating sheet should be placed under vegetation and vegetation struck ca 5 times in rapid succession with the beating stick.Ê Immediately search sheet for active specimens (weevils and ants only) and collect using aspirator, featherweight forceps, or by hand. ÊCollect into vial of alcohol. Once active specimens have been collected, search again for any cryptic and non-active specimens and collect as before.Ê Shake sheet to discard plant debris and repeat process.Ê Continue beating, working along trail for a 2-hour time period.Ê This constitutes one sample. If rain starts, stop beating.Ê In general, beating in rain is not highly productive.Ê
Many weevils live on dead vegetation, especially tangles of vines and twigs.Ê Special attention should be paid to fruiting and flowering plants as these plant parts are often preferred by weevils for oviposition.Ê Learn what seem to be preferred or the most productive microhabitats.
After lunch break, people will switch roles and repeat morning sampling. Back at the base camp, check all vials that have a high specimen to liquid ratio and change etoh.
Day 3. Repeat baiting and beating protocol in two sites other than transect sites. Record locality data for new sites.
Day 4. Flexible morning and early afternoon, collecting activities planned on-site. Good time for educational activities with communities. Late afternoon: harvest Malaise traps, harvest Winkler samples, pack.
Day 5. Transport to next site.
Samples are transported back to host institution and securely stored in a refrigerator or freezer until workshop and sample processing time.
Comments on Winkler extraction efficiency
At Playon de la Gloria, a lowland rainforest site, we took a series of Winkler samples and kept track of some methodological variables and their relation to number of ants extracted. There was little or no replication, so the notes reported here are anecdotal. First, the sifter used by Jesus Luna-Cozar had coarser mesh than the other LLAMA sifters. One sample was taken with the coarse-mesh sifter and three comparable samples with the fine-mesh sifters. All of them had four mesh sacks full in the Winkler bag. The coarse-mesh sifter yielded a total of 773 ants, while the numbers for the other three were 395, 429, and 509 ants. However, another sample taken with a fine-mesh sifter but with only two mesh sacks full in the Winkler bag yielded 669 ants. There may be some diminution of ant number with the finer mesh, but if so it is not dramatic and may be small compared to other variables that impact ant number (e.g. spatial variability, litter conditions). Second, five samples were hung at the field site for one day, then put back in cloth bags in the afternoon of day 1, transported to the university in San Cristobal, rehung in Winkler bags on the morning of day 2, and taken off at the end of day 3. The proportion of ants extracted in the day 2-3 period varied from 9 to 34 percent of the total 3-day sample.
From casual inspection, there was no indication that species composition differed between coarse-mesh and fine-mesh sifters or early and late fractions.
Collection codes are a critical database linking field between individual specimen data and the ecological data connected to those specimens. Collection codes need to be rendered exactly; any deviations will break the link that allows one to look up data for a specimen or specimens for a collection. Common sources of errors are extra spaces, invisible spaces added at the end of a code, mistaking numbers for letters (e.g. the number one and the letter l, the number zero and the letter o). Although not essential, it is common for the collection code to have some information content itself. For example, different parts of the code may indicate locality or method. If the code is to contain information, it is important for the code to be fixed length, such that particular character positions or ranges convey a particular kind of information, such that codes can be decomposed using Excel formulas.
Winkler/bait/beating/Malaise samples:
General form: m-c-xx-y-zz
m = method: Wa = miniWinkler, Wm = maxiWinkler, Wx = night miniWinkler, Ba = bait, Ma=Malaise, Go=beating.
c = country code: A=Mexico, B=Guatemala, C=Honduras, D=Nicaragua (ABCD used instead of MGHN because letters M, H, N easily confused when hand written)
xx = site number: 01 to 09 for each country.
y = transect number: 1 or 2 (additional locations 3 and 4 for baiting and beating).
zz = sample number: 01-50 for miniWinklers, variable number for baiting and beating samples, 01 for maxiWinklers.
General collecting: each participant maintains a personal field notebook with a collection code series. Personal collection code series are formed by participant's initials followed by a sequential number.
Data for all collection codes are maintained in field notebooks and entered in a laptop computer when possible.
Georeferencing
Each collection code must have the following georeferencing data: latitude, longitude, gps error, extent, datum. Latitude and longitude are in decimal degrees. gps error is the error indicated on the gps unit. Extent is a distance, in meters, that reflects the maximum possible distance of the collection from the point where the gps reading was taken. This is an important concept. A Malaise trap and a series of hand collections in the vicinity of the trap may have the same gps coordinates, but the Malaise trap might have an extent of 0, while the hand collections might have an extent of 100m. Maximum error is later calculated as gps error + extent. Datum is a setting on the gps unit (for a particular projection system), usually WGS84 for Central America.
An Excel file with all collection codes has the following column headings (fields):
CollectionCode
Method
Locality short or extended (a nickname allowing grouping of collections by site, or an extended locality name that would not fit on one line of a label)
Locality label (to be printed on specimen labels)
Date start (in format dd-mm-yyyy, where month is differentiated (e.g., 12 Apr 2008 or 12.iv.2008)
Dateend (used for Malaise samples and other collections where there is a range of dates)
Collector label (Collector field for printed label, same as next field for individual collectors, "LLAMA" when general project LLAMA codes)
Collectors (individual collector, or string of initials when teams participated in collection)
Elevation (m)
Latitude (decimal degrees)
Lon neg (longitude, decimal degrees, negative for western hemisphere, allowing easy pasting into Google Earth)
Longitude (decimal degrees)
Gps error (m)
Extent (m)
Max error (m)
Datum
Habitat label (for printed label)
Habitat (can be more extensive description, more than will fit on specimen label)
Source (for hand collections, additional info such as nest characteristics or host plant; 30-character maximum field length)
Notes (any additional info about collection, including later processing history)
Country
State/Province
day (a formula entry based on date field; for specimen label creation)
month (a formula entry for specimen label creation)
year (a formula entry based on date field; for specimen label creation)
Label formula (An Excel formula that produces a text string for specimen labels; see below)
Habitat label (a formula for a habitat label)
Project LLAMA is taking an approach to specimen labels that assumes that labels are generated for specific collection codes and printed on demand for particular batches of specimens. This is in contrast to using a generalized label for a large area, with a large batch printed all at once.
Specimen labels are of the form
MEXICO, Chiapas:
2km SE Custepec
15.72298Ê-92.94493 ±50m
1650m, 18May2008
LLAMA#Ba-A-01-1-01
The third line of the label is latitude and longitude in decimal degrees to five decimal places regardless of error, followed by the maximum error (GPS error + extent). It is important to have the error term on the label. The fourth line is the elevation, followed by date. The collection code is printed as part of the primary label, so that georeferencing data are sure to be associated with a particular collection. Important Note: the second line is the generalized locality name, which may apply to many different gps points. Thus labels may look very similar yet differ in gps point. Thus it is important to associate a collection code with a label.
[Prior to August 2009 a different format was used, with ¡N and ¡W notation, and maximum error on the fourth line. Any specimens mounted and labeled prior to this have the older label style.]
Secondary habitat labels are of the form
Cloud forest, ex
sifted leaf litter
Collaborators can download the Excel file and use the data to construct labels as they wish, but one method is as follows:
Another option is to download this Word file (click here) that has a single copy of the label for each collection code. You can search for the collection code of interest and copy the label text.
In the Longino lab, labels are printed on an Epson Photo R1800 inkjet printer, using the recommended Epson inks. Paper is Epson Ultra Premium Presentation Paper Matte. These materials are designed to have good archival properties for professional photographers. In addition to being suitable for pin labels, the ink is not alcohol-soluble and seems to hold up well for fluid storage.
Sites are in the order they were visited.
Cerro Pinal—n (high elevation): Cerro Pinal—n is located within the Reserva de la Biosfera Sierra de las Minas above Finca Trinidad. Sampling was carried out in an extensive area of cloud forest, at about 2500m elevation. The forest had abundant oak and pine, and in some areas bamboo was abundant.
Biotopo El Quetzal (mid elevation): Biotopo Quetzal (also known as Biotopo Mario Dary Rivera) is a reserve located near the city of Purulh‡ west of Sierra de las Minas.ÊThe habitat is diverse mature cloud forest. The transects were taken in the vicinity of the Reserve buildings, at around 1700m, and additional sampling was done along a trail leading up to nearly 2000m.
La Uni—n (mid elevation): La Uni—n is an area of coffee farms and extensive cloud forest, near the Honduran border south of Lago Izabal (on the south side of the Motagua fault). The forest has abundant oaks and tree ferns. Sampling was carried out between 1400-1600m.
Monta–a Chiclera (low elevation): Monta–a Chiclera is a municipal park of about 1000ha that protects the water source for the nearby town of Morales. It has second growth and mature lowland forest. The forest has a more seasonal aspect than the wet lowland forest of the Lacand—n in Chiapas or the coastal lowland forests of Honduras, Nicaragua, and Costa Rica. There were relatively few broadleaf monocots: Heliconia, Marantaceae, Costus, etc. The two transects were located in the mature forest just behind the water reservoirs on the southern edge of the reserve, around 200m elevation.
Cerro Cahu’ and Tikal (low elevation): Cerro Cahu’ is a small reserve on some low hills on the north side of Lago Peten Itza. The reserve is an island of mature moist forest habitat, surrounded by rural development. About 25km north of the lake is the edge of Tikal National Park, which is an extensive area of mature forest. The forests are somewhat like Monta–a Chiclera, seasonally dry, with few broadleaf herbaceous monocots. One transect was on Cerro Cahu’, the other in the center of Tikal National Park, just north of the famous Mayan ruins. Elevations were generally below 300m.
Machaquil‡ (low elevation): The Machaquil‡ forest was on private land, just north of Poptœn. It was patchy forest on small limestone hills, in an area of karst topography. Although the forest had the generally seasonal aspect of Monta–a Chiclera and Tikal, it seemed somewhat wetter than either of those sites, with more Heliconia, Marantaceae, etc.
Cerro Santiago (high elevation): Parque Ecol—gico Pino Dulce is a privately owned park between Mataquesquintla and Jalapa (on Cerro Santiago).Ê It is located on the side of a local peak and has trails that pass through primary and secondary cloud forest. One transect was done on a high ridge around 2600m, the other on a lower ravine slope around 2300m.
Cerro Carmona (high elevation): This is a cloud forest area immediately adjacent to Antigua. It is a coffee farm and adjacent cloud forest reserve owned by a family that is very interested in conservation and developing the reserve for Ecotourism. One transect was in fairly mature second growth cloud forest at 2150m, the other in oak forest at 2350m.
Atitl‡n (mid elevation): Refugio Quetzal is a small research station owned by Universidad del Valle.ÊIt is located at 1600 m on the southern slope of Volc‡n Atitl‡n above Finca Panam‡. The station has developed plantation land below it and mature diverse cloud forest above it. The two transects were in mature cloud forest near the station building. Some additional sampling to 2400m was carried out along a trail that continues up the volcano slope.
Sampling protocols were similar to 2008 with the following modifications.
MiniWinkler transects were carried out by teams of 4 instead of teams of 2. MaxiWinklers were not taken simultaneously. When a team finished a transect, they took a few maxiWinkler samples on the way back. More effort was made to take maxiWinkler samples at sites other than the transect sites.
Michael Branstetter added a project contrasting night versus day miniWinklers. At one low site, one mid site, and one high site, the LLAMA crew sampled a transect of 15 miniWinkler samples at night, in the vicinity of one of the standard day transects.
See notes for 2008. This year's collection codes have "B" as the second element, after the method.
See notes for 2008.
This Word file (click here) that has a single copy of the label for each collection code. You can search for the collection code of interest and copy the label text.