This "fundamental computing problem" is the problem of being able to quickly and efficiently pick out objects that are similar to one another from very large data sets.

For example, Google needs to be able to pick out search results that are similar to what you typed—and it needs to do this very quickly. Likewise, Netflix needs to compare your watch history to that of all its other users so that it can make recommendations for what else you might like.

These and other similar problems pop up all over the place in your everyday life. The process of solving them is called a "similarity search" and is of intense interest to computer scientists who want to make these searches faster, more efficient, and better quality.

In this experiment, the authors injected normal Yob transcripts into normal embryos, along with a fluorescent marker to let them know that injection was successful, and allowed the embryos to grow. Once at the larval stage, they could see the fluorescent injection marker and sorted the larvae based on expression of the marker. At the pupal state, the authors were able to sex the mosquitoes.

Based on what the authors found by analyzing other species, they hypothesize that Yob actually codes for a protein. They made new lab-generated transcripts of Yob with sequence changes that would result in no protein product (either removing the start codon, or adding in a premature stop codon) and transfected those transcripts in to cells to monitor dsx splicing.

The authors isolated sequences from closely related mosquito species and compared similarities and differences in those sequences computationally.

The authors made mRNA in the lab. The mRNA has the same sequence of what they think is the shortest, but functional Yob transcript. They are testing whether Yob has a direct effect on dsx.

The scientists artificially inserted the gene-reading Yob into the embryos of Anopheles gambiae.

As seen in Figure 1, they are analyzing Yob mRNA through use of a gel, not sequencing, in this experiment. The authors find that transcription of this male factor begins very early after fertilization, at the same time as other genetic markers of zygotic expression.

Because mosquito embryos look identical whether male or female, the authors had to use molecular techniques to find out the sex of each embryo. The embryos were separated based on their sexual identity (male or female) and mRNA was extracted and sequenced from the collective male or female embryos.

This experiment is an extension of the frameshift experiment. If the bases are read in triplets, then you'd expect that four or five extra bases would ruin the code. However, six extra bases would still encode the amino acids for the native protein (with two extra amino acid residues).

The authors performed a power analysis study, which allows researchers to determine the sample size required to detect an effect, due to the large number of samples with broad characteristics. They first calculated the number of samples needed to determine a difference in microbiota diversity when the cause is unknown. To do so, they need to determine the effect size (the minimum deviation that is considered significant), the significance level (the probability of determining that a condition is true given that it is false), and the power (the probability of determining that a condition is false given that it is true).

Read more: http://www.biostathandbook.com/power.html

Genetic sequences of the modern and ancient dogs used in this study are publicly available. Check out their data on Dryad!

A method of estimating the time at which populations had a common ancestor, based on their genetic differences and similarities.

A technique that looks at the pattern of DNA sequence changes (mutations) in multiple individuals, focusing on the most recent common ancestor of any two sequences. It can provide information about the timing of shared ancestry, population sizes, population splits, and migration rates.

Certain bases that are always present in dogs were covered by sequencing 28 times, so the whole genome is said to have been covered 28 times.

Acronym for standard error of the mean, a measure that represents how far the mean of a sample is from the estimated true population mean. It is a good estimate of how accurately your mean reflects the true population.

To learn more about SEM: https://www.khanacademy.org/math/ap-statistics/sampling-distribution-ap/sampling-distribution-mean/v/standard-error-of-the-mean

An analysis of covariance is used to see how two independent variables change together.

Here, the authors used it to determine whether the response to a reward or devaluated stimuli was due to habitual learning, rather than external factors such as outcome-action knowledge, working-memory/inhibition, or the ability to learn how to discriminate during the training phase of the test.

To learn more about analysis of covariance: http://www.lehigh.edu/~wh02/ancova.html

The final thing the researchers wanted to test was whether artificial activation of the NPF circuit was inherently rewarding. To do this, they once again relied on the dTRPA1 proteins, which can activate neurons at high temperatures. The researchers inserted this protein into neurons that were part of the NPF circuit. They then exposed the flies to one smell while keeping a high temperature (thus activating the NPF neurons). This smell became associated with a state of high NPF circuit activity. Conversely, they exposed the flies to the second smell while keeping a low temperature. This smell became associated with a state of normal/low NPF circuit activity. Finally, they placed the flies in a Y maze (as before) and checked to see if the flies would prefer the prong containing the NPF-activity paired odor, or the unpaired odor.

The researchers used time spent in the two prongs of the Y maze to develop a preference score. Per this score, a positive value would indicate that the flies preferred the odor that was associated with one of the three activities (mating, alcohol exposure, or activation of the NPF circuit). Conversely, a negative score would indicate that the flies preferred the odor that was not associated with these activities. A score of zero would indicate no obvious preference between the two.

So far, the researchers have shown that changes in NPF levels correlate with both sexual experience and alcohol preference. Specifically, a more active sexual history is correlated with an increase in NPF levels, which is correlated with a decrease in preference for alcohol (and vice versa). However, one of the most important rules to remember in science is that correlation does not equal causation. The researchers thus decided to follow up with a direct genetic manipulation of NPF in order to establish a causal relationship.

As a final follow-up experiment, the researchers took a group of sexually rejected males and split them up into two. One subgroup was left alone, and then tested for alcohol preference. The other subgroup was exposed to virgin females just before testing. As such, these males did in fact get an opportunity to mate before undergoing the alcohol preference test. If the first subgroup continued to prefer alcohol, but this preference were to disappear for the second group (who initially experienced rejection, but then were able to mate just before testing), it would indicate that the main cause for the alcohol preference was indeed sexual deprivation.

For the second follow-up experiment, the researchers wanted to investigate the possibility that it was not actually the lack of sex, but rather the experience of rejection by a female that was driving the virgin males to prefer alcohol.

A redundant covariate is a factor that either has the same value for every sample or multiple factors that are highly correlated and can be used as substitutes for one another. The researchers removed redundant covariates from their study because they can hamper interpretability of the results.

This is a statistical technique used to describe linear relationships between components of dependent variables that can be explained by a set of independent variables. Here, the independent variables were the 69 covariates and the dependent variable was the microbiota variation in each sample.

Stands for Principal Coordinate Analysis, a statistical technique used to visualize similarities between variables by representing the variance in as few axes/dimensions as possible. The arrows show the direction of the original variables and the angle between arrows gives a rough estimate of the correlation between features.

Almost all bacteria have a copy of the 16S rRNA gene. Comparing their sequences is a common tool to identify and differentiate bacteria at the species level.

Learn more about how bacteria are identified with this resource from iBiology: https://www.ibiology.org/microbiology/genome-sequencing/

For the fly, the data is organized into a higher dimensional space. For LSH, it is organized in a lower dimensional space. But if the algorithms work effectively, the data should still be arranged so that similar features are near one another.

Instead of feature vectors, the algorithms arrange items in hashes or "tags" (just like the fly brain uses a tag to represent an odor, the algorithms use a tag to represent a specific image or word). Finding the hashes that are closest to each other in this new m-dimensional space should reveal the images/words that are most similar to each other (again, if the algorithm works correctly).

These nearby hashes are called "predicted nearest neighbors" because they predict which items in the data set are the most similar.

In a mathematical proof, the authors took a matrix of input odors and multiplied it by a vector of projection values to create a new matrix of Kenyon Cell values. In this new matrix, the relative positions of the input odors were approximately the same as they were in the original matrix.

This shows that the fly algorithm preserves the distance between input points when sending information from PNs to KCs.

In bright field microscopy, an image is produced by bouncing light off of an object and the image is magnified by lenses. In scanning electron microscopy, an image is produced by bouncing electrons off of an object and the image is interpreted and produced by a computer. By using electrons instead of light, researchers can observe incredibly fine detail in the image created.

Researchers look at a sample through a bright field microscope to look for evidence of cell morphology that could be basidiomycetes.

To take a measurement of a sample, usually meaning a biochemical or chemical measurement of a biological sample; in this case, basidiomycete yeasts.

In a biological sample that contains more than one organism, before gene expression can be evaluated, the researchers must identify which genes belong to which organism. To do this, gene expression data is mapped onto the genomes of the organisms that are known in the mixture, and genes are sorted, or binned, into groups by the genomes they match. This way the gene expression data for each known organism can be determined.

The authors knew that the difference between these two lichen species was poorly understood. One species produces vulpinic acid, and the other does not. Phylogenetic analysis could not identify differences in the symbionts of the two species. The authors decided to test if differential gene expression of a gene shared between the species might be the cause of the phenotypic difference.

Using genetic engineering, the authors isolated the protein strand the two organisms had in common and exposed it to a PET film without any microbes present.

The authors collected samples from different environments where PET plastic waste was likely to be found. In this case they chose to sample near a PET recycling plant. The plant specifically recycled high-crystallinity PET. They took some dirt, water, and sludge and put the samples in an environment including low-crystallinity PET. Low crystallinity PET is more disorganized and easier to break down than polymers with higher crystallinity.

It was ultimately one of the samples of sediment that grew bacteria. When looking at the bacteria under a microscope, the researchers were able to see that it was growing on a thin layer of PET plastic.

The authors used published data of proteins having similar activity as the ISF6_4831 enzyme and analyzed it using statistical tests to select three enzymes that divergently evolved—meaning that the enzymes evolved from a common ancestor and accumulated enough differences to result in the formation of a new species.

This technique is used in calculations for complicated systems where changes in one variable influences many parts of the calculation. In this technique many calculations are run with each calculation having different beginning conditions. The end results are then analyzed to look at the range of possible outcomes. The error bars in parentheses in Figure 4 are determined by a Monte Carlo simulation.

See more about Monte Carlo simulations from the Massachusetts Institute of Technology: http://news.mit.edu/2010/exp-monte-carlo-0517

R-squared (\(R^2\)) values represent how close the data points are to the predictions from a model. An \(R^2\) value of 1 means that all of the data points lie on the prediction from the model. In this case, the researchers are using the solid line that represents perfect agreement between the box model calculation and the experimental data. An \(R^2\) value of 1 in this case would mean that the box model calculations and the experimental data are in perfect agreement. In general, the closer the \(R^2\) value is to 1, the closer the model and data match.

Without the inclusion of VCPs in the model calculation the \(R^2\) value was 0.59 (Panel A of Figure 3). With the inclusion of VCPs into the model the \(R^2\) value improved to 0.94 (Panel B of Figure 3).

The air in Los Angeles, California has ozone (\(O_3\)) levels that are higher than those recommended by the Environmental Protection Agency. The amounts and types of VOCs present influences ozone levels.

Comparisons are made between experimental data and model calculations. The model calculations are influenced by what molecules are emitted, how those emissions chemically react, and how wind and other variables cause emissions to move from one place to another. The calculations use techniques from other researchers (including this paper's co-authors Joost de Gouw and Si-Wan Kim) to take into account chemical reactions and movement of the emissions from one place to another.

The authors then make different predictions with their model assuming different types of chemical emissions (fossil fuels and/or VCPs).

Box model calculations were carried out under various conditions. The authors can make various assumptions about what sources are emitting VOCs into the air (mobile sources or VCPs). The results of the calculations under these various conditions are then compared to experimental data. This allows the researchers to examine what sources are important contributors to VOCs in air.

The authors define a series of constraints for their simulation. The scientific model assumes facts about the phenomenon that make the problem easier to solve. In this case, the authors assume among others that the pore structure is homogeneous, that the vapor pressure is the same for the MOF and the condenser, and that the diffusivity of the crystal does not vary.

To simplify the experiments, the author maintain the condenser at a temperature cooler than the surrounding temperature. This results in the condenser being cooler than the walls of the box, so that all the water condenses on the condenser and can be easily collected. In practice, ambient temperature is enough to harvest water without the need for additional cooling of the condenser.

The authors simulate natural sunlight in the laboratory. The air mass is the ratio of the path length that sunlight has to travel through the atmosphere to the vertical path length (when sun is at the highest point in the sky, which corresponds to the shortest path length). An air mass of 1.5 is the condition most commonly use to quantify the efficiency of solar devices.

In order to efficiently harvest water, a large difference of temperature between the MOF and the condenser is needed. In their experiment, the authors set the temperature at 35°C in order to be able to condense the vapor with a condenser at 10°C.

The authors performed a simulation in conditions reflecting an accurate, real-life use of the MOF. The results show an increase in RH upon water desorption, followed by return to the initial 20% value upon water adsorption.

The diameter of MOF-801 was determined by a method called scanning electron microscopy. The surface of the MOF is scanned with a beam of electrons which gives information about the surface.

See figure S5 in the Supplementary Materials.

The authors developed a computational model to simulate the behavior of the MOF material. Through this approach, they can study the influence of the material parameters on the water harvesting properties and then confirm their result experimentally. Their goal is to screen a large number of parameters computationally to optimize the system.

The author first quantified the water harvesting capacity of the MOF. For this purpose, they use 20% relative humidity for their measurements, as it is representative of the low humidity in dry regions of the world where an efficient water-harvesting method is most needed.

The design of the prototype results in a slower desorption process than was considered in the initial calculations. The authors modified the parameters of their simulation to take this into account and have a model reflecting the experimental conditions.

To demonstrate the utility of their material in real conditions, the authors built a small device containing the MOF and a condenser and tested it outdoors.

The MOF is pressed onto a substrate made of copper metal in order to have a material that conducts heat efficiently and that can cover a large surface.

Like PGLS, this method examines the relationship between two or more variables. However, this method does not account for phylogeny. OLS regression was used to compare with PGLS results to see if it would produce similar results.

To test whether climate change has impacted species ranges, the authors had to measure, for each species, the latitudes of the northern and southern range limits (and the distance between these), the warmest and coolest temperature the species were observed, and the mean elevation of the species.

These measurements are based on the average of the five most relevant species observations in each case. For example, to find the northern range limit, the authors took the average of the five northernmost observations of the species, while to find the coolest temperatures within the species' range, they took the average of the five coolest points that the species was observed at. This helps avoid skewed results from a signal sighting.

Check out the Supplemental Materials for more information on how the authors took all their measurements.

Glial fibrillary acidic protein (GFAP) is an intermediate filament (cytoskeletal component) protein expressed by glial cells including astrocytes. Glial cells are non-neuronal cells within the central and peripheral nervous systems whose roles include providing support and protection for neurons. When a gene like ChR2 is under the control of a GFAP promoter, only those cells that express the transcription factor for GFAP will express ChR2.

To test whether optogenetics can be used to dissect neural circuits, the authors utilized an established model of induced Parkinson's disease. 6-OHDA is a synthetic neurotoxin that selectively destroys dopaminergic and noradrenergic neurons in the brain. When injected on one side, it can cause side-biased motor deficits. The right medial forebrain bundle is a neural pathway that passes through the hypothalamus and basal forebrain and contains a high number of dopaminergic neurons.

This is a mathematical statement that shows that you can take a small number of points in a high-dimensional space and transfer them to a lower-dimensional space while (very nearly) maintaining the distance between all the points.

You could think of this like taking all the points on a sculpture and projecting them onto a piece of paper while keeping every point in its proper position relative to all the other points on the sculpture.

Here, the authors use chemicals (ethanol, methanol, and dimethyl sulfide) to demonstrate similarity tagging/hashing.

As stated earlier in the article, the authors could only fairly compare the LSH and fly algorithms if each used the same number of mathematical operations.

They determined that if they used 20k random projections in the fly algorithm (where k is the length of the output tag, or hash) then the total number of operations for the fly and LSH algorithms would be equal.

For more detail, see the second paragraph under "Materials and Methods" in the Supplementary Materials document.

From the 10,000 vectors, the authors randomly chose 1,000 of them to be query inputs (images or words that they would test against the rest of the feature vectors to find which ones represent similar images or words).

They chose 10,000 vectors of length d from each data set. Each vector represented an image (SIFT, MNIST), or word (GLOVE).

This is a method of measuring how well an algorithm performed. The more overlap between the two lists, the more similar they are. If they are exactly the same, this means that the algorithm performed perfectly.

For example, it might mean that it found the closest match to all query images, or that it correctly found all the correct features that match part of an image (see picture below).

Each feature vector contains a row of values representing features of an image or word (depending on the data set). Because the data is arranged so that similar features are near each other, measuring the straight-line (Euclidean) distance between two vectors allows the authors to determine the similarity between them.

That is, the closer the vectors, the more similar their features. The vectors closest to one another are the "true nearest neighbors," or the images/words in the data set most similar to one another.

The Akaike information criterion (AIC) is a statistic that measures whether some models are more or less informative than others. This means that a model with a lower score for AIC is more informative than a competing statistical model and may tell us something that is meaningful in a biological sense.

Like many other aspects of statistics, we have to be careful about blindly believing test results, so we apply our experience as scientists to make sure that numbers like the AIC score actually make biological sense!

Click here for a more in-depth view of what AIC is, how it is calculated, and what it can (and can't) do.

Phylogenetic Generalized Least Squares (PGLS) is a method that accounts for phylogenetic relationships within a group of species (phylogeny reveals how closely related species are to each other). This method will determine if variables of interest are closely related.

In this table, models of trait evolution indicated whether traits are ancestral and kept over evolutionary time. Species’ upper thermal limit (one of the traits studies here) was found to be shared by close relatives. The authors pointed out in this paper that niche conservatism (i.e. the tendency for species to keep their ancestral traits) could explain such findings.

Without also testing competing theories (land-use and pesticides), the authors would not be able to say how important the effects of climate change were in influencing range limits. Remember that the scientific method does not allow us to definitively prove a theory, only add to the evidence that one hypothesis is more likely than the alternatives.

Because climate change is not the only possible explanation for changes in species' latitudinal and thermal range limits, the authors also tested other competing theories, like land-use changes and pesticides.

To accurately target the subthalamic nucleus (STN), the authors used extracellular recordings to determine accurate position within the deep brain structure since the STN has a characteristic firing pattern. Guided by this method, accurate injection of opsin vectors to the region was accomplished.

The authors used satellite imagery and characteristic wind and moisture patterns in order to gauge the intensity of storms and hurricanes.

This method of gauging hurricane intensity has been tested against aircraft sensor measurements and does a solid job of accurately recognizing hurricane characteristics.

These authors used satellite data from 1970 to 2004 in order to examine cyclones and hurricanes in every tropical ocean basin.

They examined: 1) number of storms and hurricanes 2) number of storm days 3) hurricane intensities

Most of the data examined came in the form of best track data archived in hurricane warning centers.

A method used to reinforce an association with a certain stimulus. Unlike classical conditioning, instrumental learning (otherwise known as operant conditioning) is active and involves a person performing behaviors that are positively or negatively reinforced. Reinforcement is also known as learning.

The two instrumental tasks used in this study were reward and avoidance learning. Participants learned that using a pedal would allow them to avoid being shocked when presented with pictures that they had learned to associate with a shock.

A test for habitual behavior. Certain stimuli that were previously associated with a reward are devalued, meaning the reward is removed.

Once the reward is removed, responding to a stimulus no longer makes sense. If people continue to respond, their response is considered habitual.

Skin conductance measures the activity of certain sweat glands, which become more active during the avoidance response (resulting in a higher skin conductance).

It was measured 0.5 to 5 seconds after each stimulus was presented. Conductance allowed the authors to measure how the participants respond to seeing the stimulus before reacting to it and to make sure they had learned to fear the stimuli associated with a shock. This way, they knew that if participants performed poorly in the task, it was not because they did not learn to fear the stimuli.

The authors based early hominin stature estimates on the length of intact fossil femurs, reconstructed femurs, and femur head diameters.

The authors used average footprint length when estimating stature and body mass. They did so to minimize overestimates and underestimates of footprint size and to compensate for poor preservation.

The authors used a regression equation based on walking pace as it relates to the size of the footprint and body mass. Australopithecus afarensis proportions were used rather than those of modern humans.

The authors also estimated the stature of the Laetoli trackmakers using Australopithecus afarensis foot:stature ratios.

The authors used two different methods to estimate the stature of the hominin trackmakers. One used modern Homo sapiens body proportions. This was considered to be less accurate since the body proportions of the hominids were likely different from those of modern humans.

The authors gathered the following data on tracks from both G and S.

The authors used the same data gathering and visualization techniques on the fiberglass casts as they had used on the tracks at S.

The authors checked their measurements and used STAR*NET software to combine their conventional observations with GPS vectors. They used the leveling observations the software provided to make 3D adjustments in the footprint images.

Each test-pit was surveyed and photographed in order to determine its relationship to the other test-pits and to Site G.

The authors carefully measured distances between targets so that the location and depth of each of the modeled footprints could be precisely determined.

The authors made 3D models of selected footprints and sets of footprints. Models of some of the footprints were not made due to their poor state of preservation.

Here, the authors explains how they ran different scenarios in the simulation system with the data they had collected in the field, making for more realistic results.

These data collected from 1963-2012 wereinputted into the DAYCENT system to simulate conditions during the next 100 years (from 2000-2100).

This table shows the two different sites, Taylor Slough (TS) and Shark River Slough (SRS) along with their exact location using the coordinate system, characteristics of each site, such as root:shoot ratio, soil composition, depth of roots, Nitrogen deposition along with the amount of Carbon found at each site.

In this experiment, grids were designated over the island and 30 x 30 meter plots were established to sample within. In these plots the total soil nitrogen, phosphorous, amount of stable nitrogen isotope, and percentage nitrogen were sampled across grasses, forbs, dipterans, arachnids, passerines, and mollusks to determine how much of these nutrients were derived from the ocean. Organisms and soil that derive their nitrogen from a local source will have fewer amounts of nitrogen isotopes than those that get their nitrogen from higher trophic levels as is shown in Figure 3.

In this experiment, nutrients were added to a community that represented the fox-infested islands which increased the grass biomass and also created a more grass dominated environment overall.

The nutrient additions were meant to represent the spread of guano by the seabirds. In the fox-infested islands the seabird populations were scarce, which decreased guano accumulation. Therefore, the added nutrients represented the guano production by seabirds in the absence of seabird on the island that is fox-infested.

These nutrient additions are similar to the nutrient states of fox-free islands and show what the changes in ecosystem could have been like during the introduction of foxes.

In another experiment, they tried to mimic the disturbance in soil of seabirds burrowing, this negatively impacted the grass biomass and it was shown that fertilization by itself had the greatest positive effect on grass biomass and distribution. This helps explain the higher incidence of grasses in the fox-free islands.

This figure shows the islands that were studied and their relative position on Earth. There are 9 fox-infested areas shown in red and 9 fox-free areas shown in blue, the sample size of islands is the same so the results cannot be attributed to the sample size difference. The yellow dot is the location of fertilization experiments.

This is very broadly what the experiment seeks to discover through looking at fox-infested and fox-free islands. This experiment is designed to assess effects of sea bird densities on soil nutrients and composition, plant abundance, primary productivity, and the flow of nutrients from lower organisms to higher ones (i.e. from plants to herbivores to carnivores.)

It would be more accurate to say that the models used to describe physics had to be modified.

As physics progressed, physicists revised their theories in light of new evidence and knowledge (and they continue to do so today). Here, Einstein describes how the special theory of relativity came about, in part, to solve the problem of classical mechanics' compatibility with electromagnetic theory.

Modern physicists consider classical mechanics an approximate theory that is useful for the study of non-quantum mechanical, low-energy particles in weak gravitational fields. It is usually the starting point for students learning physics.

Different genetic sequences melt at slightly different rates. Therefore, they can be analyzed using these melting point curves, where fluorescence is plotted on the y-axis and temperature on the x-axis.

If the DNA strand is composed of many guanine and cytosine bases, it will take a higher temperature to break the hydrogen bonds apart as opposed to adenine and thymine bases. This is why the E2 clade fluoresced at a higher temperature than the other clades.

Denaturing Gradient Gel Electrophoresis (DGGE) has been used, along with yeast cultures, to analyze saliva samples of 24 adults to study the bacteria present. This study showed ample variation of bacteria among individuals. The conclusion obtained from this experiment was that there is a lack of association between yeasts and bacterial DGGE fingerprint clusters in saliva, implying a significant ecological specificity.

This study (experiment) wanted to identify the effectiveness of high-resolution melting for the genotype (genetic makeup) of the Symbiodinium. Symbiodinium are single-celled algae that can be found in the endoderm (innermost layer of cells or tissue of an embryo in early development) of tropical cnidarians such as corals, sea anemones, and jellyfish.

A technique that detects mutations, and differences in DNA samples. HRM has been used to target various microbial communities on tadpole intestines and feces. In this study, HRS targeted a short amplicon (piece of DNA or RNA that is the source of amplification or replication events) of the 16S rRNA gene, which was the sequence being tested. Along with the HRM, gel electrophoreses and DNA sequencing were also used in order to study the results more closely. All three methods revealed several types of bacteria living in a tadpole's intestines and feces.

Since much of the calculations were done using estimations, the authors could not provide a range of values that's likely to encompass the true value. For example, when scientists say this value falls within the 95% confidence interval it means that the intervals contain the true value 95 percent of the time and fail to contain the true value the other 5 percent of the time.

In order to determine the pollination efficiency of the insects, the authors tracked how often a specific visitor groups arrived and how long each individual stayed. Then the authors calculated an “efficiency” by estimating pollen on the visitors’ mouthparts. This was calculated as the average number of pollen grains per individual visitor of each group.

The reason behind this technique is to determine what species the carriers favored.

The author monitored when the pollinators came around and interacted with the plants. Reason behind this is to determine how frequently the flower will have visitors. Another reason why the author monitored the plants, was to determine the period of time when the flower had the most visitors. With both of these factors being monitored, the author can determine more or less how frequently these plants can have their pollen spread throughout an area.

This experiment isolated the genome of an individual C. argentata sample using the same method described in the previous experiment (DNeasy). The DNA sample obtained from this group was then taken to the Georgia Genomics Facility at the University of Georgia to develop microsatellite loci which are single sequence repeats that allows researchers to see any variability in the genome of this specific plant compared to the others. The researchers then used an Illumina HiSeq 2000 to see the microsatellite markers. The process through which they obtained these microsatellite loci is not described in the paper; however, the significance of this step when comparing different samples of DNA is described in the video below: https://www.youtube.com/watch?v=9bEAJYnVVBA

Here, the authors are hypothesizing that the EOD pattern incorporated into electrosensory nervous system of the electric fish uses the information of size, shape, and distance of the objects in an algorithm to process and relay information to the electric fish brain.

The authors here talk about the premise of their experiment and how they want to take all the data that the team has collected visually and through pattern tracking and turn it into data that has numbers. That is why in the chart that follows there is system mapping for EOD of both "wave" and "pulse" fish.

Table 1 shows data indicating linear relationships between variables that changed in an evolving population based on the variability of the environment. The P column explains whether the data is statistically significant, meaning it can be concluded as a relationship between data and not just due to random chance. The two values observed for each variable are Y intercept (which indicates whether the trait changed from the original population) and the slope (which indicates how much the trait changed). These values are calculated as P values to determine if they are significant.

Figure 3 shows the initial and final genotype compositions as well as trait changes among the different conditions (A. continuous, B. gap size of four times the mean dispersal rate, C. gap size of eight times the dispersal rate, and D. gap size of 12 times the dispersal rate). The central pie chart found in each of the four graphs shows the equal frequency of genotypes in the founding population, while the other pie charts are used to represent the genotypic make-up of the top 10 leading individuals after 6 generations of spreading within the different conditions. The placement of the pie charts is based on rankings of three traits: the competitive ability of the plant (dominance of the plant in ways that do not directly affect it's ability to spread offspring), dispersal (the average distance of the farthest dispersed seed from the plant it was dispersed from), and the height of the plant.

These values show that even as the gap size increases, the trait values hardly increased at all. Similarly, as can be seen by the color of the pinwheels, the genotype composition change between the different gap sizes was rather consistent. The only outlier is the continuous condition that has random trait values and genotype compositions.

In the first experiment, the authors simulated invasion by planting 14 genotypes on one side of the pot array. The plants were allowed to proceed for several generations to follow evolution in A. thaliana. Habitat patchiness was tested by putting space between the pots at varying distances.

Notice that each bar in the figure has small capped lines that extend above and below the bar. These lines represent error bars for the various VOC emission factors. Error bars reflect the experimental uncertainty in determining VOC emission factors. The error bars in this figure are 95% confidence intervals, meaning that we are 95% certain that the true value lies somewhere in between the top and bottom of the error bar. Any measurement made has uncertainty associated with it.

In scientific communication it is important to give the value that was found and the uncertainty in that value. We typically talk about uncertainty as being the precision of the measurement, where a smaller uncertainty is a more precise measurement and a larger uncertainty is a less precise measurement.

Economic data was used to estimate the mass of various chemical products used in the United States. You can see how the estimates were derived in Tables S2 and S3 of the supporting information.

Tables S2 and S3 can be found on pages 23 and 24 of the Supplementary Materials document

Every emission source of VOCs emits different compounds. Many databases of the different emission profiles exist. The emission profiles allow the researchers to determine what activities (driving a car, painting a house, spraying perfume, etc.) lead to the presence of different VOCs in the air. The researchers use these profiles to differentiate between sources that come from fossil fuels (e.g. driving a car) and those that come from VCPs (e.g. painting or perfume).

Previous studies have measured volatile organic compounds in outdoor and indoor air around Los Angeles, California. This study is using that data to see how it agrees with model calculations when accounting for volatile organic compounds from different sources. The video below gives an overview of how atmospheric modeling helps us understand what is occurring.

https://youtu.be/rOClF26-Bo8?t=1m

The study looks at two different environments (indoor air and outdoor air) that can influence each other.

See Figure S4 on page 18 from the Supplementary Materials document

Emissions from volatile chemical products was determined by looking at the composition of chemicals in a particular product and how readily those chemicals evaporate into the air. The authors compiled information from various other researchers to perform these calculations.

United States government data was compiled to look at the amount of oil and natural gas supplied and what these sources are used for. See Table S1 of the supporting information to see the breakdown.

Table S1 can be found on page 22 of the Supplementary Materials document

The authors use different calculations to determine VOC emissions from mobile sources and VCPs. This is due to the different end uses. Oil and natural gas used for mobile sources undergo combustion that causes much of the carbon to be emitted as carbon dioxide and not VOC. VCPs are primarily used directly and remain as intact organic molecules.

Many experiments from other researchers are utilized in this study to compare model calculations with experimental data.

The authors in this study utilize all of these different sources of information to build a better understanding of the sources of air pollution.

After collecting the samples the authors provided the samples with water, an appropriate temperature, and a food source of PET film. They later used microscopy to observe what samples were able to use the PET film as an energy source.

The authors took the DNA from Ideonella sakaiensis and created a list of the order of the DNA bases A, G, C, and T. Next, they used a computer program to compare the sequence to sequences from other organisms that were already known to degrade PET.

A lab technique that isolates a single species of bacteria from a sample with more than one species of bacteria.

These results suggest that bumble bee tongue lengths are decreasing and corolla tube lengths are also decreasing in response to climate change.

From 2012-2014, researchers analyzed the change of flower production of six bumble bee host plants over time at a altitude of 400m. This is where the bumble bees most commonly forage. The flower abundance was measured as flowers per square meter across five different habitats. It was then compared to data from 1977-1984 to determine change over time.

Over the past 56 years, climate change has caused summer temperatures at Niwot Ridge, Mount Evans, and Pennsylvania Mountain to rise. To understand the relationship between summer temperatures and flower density, four bumble bee host species were analyzed for change in average flower density between 1977 and 2014.

The author listed four possible processes responsible for the tongue length changes in the bees studied:

1) Decreased body size over time has correlated to a shorter tongue - The authors compared the body size measurements with the tongue length measurements over time.

2) Coevolution between the flowers and bees - The tube depth of the flowers were measured and compared to the tongue length of the bees over time.

3) Competition from other bees in the same region affected tongue length - The authors compared other bee species to the bees studied and determine which species had the advantage and if these advantaged affected the other species.

4) Diminishing floral resources - The authors analyzed the effects of lower amounts of flowers due to increased temperatures on the foraging habits of bees. They then concluded if this could have impacted tongue length.

Scientists started using fluorescent tags as a natural way to accurately detect low concentration of metabolites in insects. This method proved to be sensitive and effective and most importantly stable. Being stable was an important factor in this method because if it was not stable it would start to degrade after about 15 minutes. The new method has lasted over an hour, which is in benefit of being detectable by HPLC-FD (High-Performance Light Chromatography with Fluorescent Detection).

The hypothesis is saying the mosquitoes' nutrition has an effect on "insulin sensitivity" and "juvenile hormone synthesis." Insulin is a hormone that works with the amount of glucose (sugar) in blood. A juvenile hormone is a hormone in insects that work with maturation and reproduction. So the researchers are saying that the amount of nutrition (food) the mosquitoes eat will determine the amount of these two hormones, insulin and juvenile, are produced.

Since the researchers hypothesize that how much the mosquitoes eat determines how much insulin and juvenile hormones work, this means how much insects reproduce is also affected. So the researchers are saying if the insects eat enough, they will reproduce with better conditions than if they were not eating enough. This is because the hormones that control reproduction are controlled by the insects' nutrition.

Certain variables in the frame of these barracuda, as they were being detected, was used through complicated physics and mathematics to calculate their stride lengths.

Ensuring the white muscle tissue were stimulated by the electric pulse to contract, tighten.

The time it took to hook a fish, reel it in, and brought on board was never more than 15 minutes.

The x-axis shows where along the length of each fish the contraction times were collected from, starting from the head (0%) to the tail (100%).

In statistics, 'P' is the p-value. The p-value is the probability that the null hypothesis is true, and if it's below .05 that usually means the null hypothesis (there is no correlation between the two variables being observed) is rejected as not true.

In the case of this section of the paper, a statistical test was run to see if the mean temperatures of the fishes (taken from the muscles being observed) were significantly different from each other. It was found that the probability of this was so low (below .05), they were not significantly different.

This is important because it implies that all the muscles of all the fish had close enough temperatures to each other that the researchers wouldn't have to worry about slightly different muscle temperatures making the contraction times collected in each fish faster or slower in relation to each other.

Basically, they checked that the muscle temperatures in each fish were close enough to each other that the variable temperature would be (close enough) to being a "constant" in all the tested subjects in the experiment.

This refers primarily to effects of strong and weak water currents, as discussed in the last few annotations.

Sailfish, barracuda, little tunny, and dorado are four marine predatory fish species known for their extremely high swimming speeds.

The researchers decided to reconsider how accurate the previously predicted estimates of each of those species' maximum swimming speeds truly is, hypothesizing that they may be over-estimations.

They did this by using a different method than that used to make the estimates they're investigating. Their method was to measure the time it takes for every contraction of swimming muscle (that doesn't use oxygen) when it twitches.

The Mantel tests allowed the authors to determine the amount of difference between species. The test uses estimated distances to measure difference.

In this case, it is testing chemical differences between species. The results indicated Piper species that are more chemically different are more likely to exist within the same environmental patch.

A Mantel test measures the correlation between two matrices. In this case the Mantel test was used to estimate the correlation between species phylogenetic distance and their co-occurrence.

Using this test, the researchers were able to statistically verify that there is a negative relationship between how closely related two species are their co-occurrence.

Pianka's Index is used for field measurements of the niches. They can measure microhabitat, resource usage, and spatial activity. The values range from 0 to 1, no resources used in common between species to complete overlap in resource use, respectively.

Gotelli's c-score measures co-occurrence between species. A binary matrix is used to represent the absence (0) or presence (1) of a species. Each column represents a site and each row represents a species. Once the c-score is calculated a high value indicates less co-occurrence and a low values indicates there is more likely to be co-occurrence.

Here, the author’s evaluated whether or not similar species tend to coexist together using the inverse nearest relative index (-NRI), which asses clusters of species, as well as the inverse nearest taxon index (-NTI), which asses cluster taxonomy of clustered species.

Along with data obtained by the chemical dendrogram the clusters were analyzed and diversity in relation to clustering was discovered.

In order to assess the dispersion of Piper species, the authors analyzed the secondary chemical composition of the plants. In plants, secondary composition are organic compounds used for plant defense. They are secondary because their absence does not cause immediate death of the plant.

The analysis of the secondary chemical composition was performed using a chromatography technique: GC-MS. This technique does not detect all secondary metabolites but it detects most of them (86% of the metabolites).

After the secondary metabolites were detected, a library was created for all species and the library was compared to references in the AMDIS (Automated Mass Spectral Deconvolution and Identification System -a mass spectra identification system). The library created was a mass spectra library.

Mass spectrometry is an analytical technique that identifies molecular weights of compounds which helps in the identification process. The particular identification method allowed the researchers to assess chemical similarities between individual plants and species.

This analysis allowed the researchers to conclude that the Piper species were more spread apart than expected when their secondary metabolites was the point of comparison.

In order to asses species coexistence, the amount of niche overlap needed to be found. Niche overlap occurs when two or more organismal units use the same resources. More overlap would mean that more species exploit each resource. This was done using Pianka's Index.

K statistic is a statistical analysis of the phylogenetic signal based off a Brownian motion metric that determines the strength of the phylogenetic signal.

Piper is a very species rich genera of flowering plants. Molecular phylogenetics has been crucial in identifying the monophyletic groups of this genus.

The researchers amplified the ITS region and the psbJ-pet A intron of each of the sampled species they collected to conduct a phylogenetic analysis. They were able to construct a phylogenetic tree of the Piper species that were found within the patches they sampled.

The authors collected data regarding chemicals produced by the plants used in their experiment using a multi-step process.

They first preserved leaves they collected within silica gel before transporting the leaves to the University of Missouri, St. Louis. There, the researchers crushed the leaves under liquid nitrogen. Next, a methanol-chloroform solution was used to extract volatile compounds along with a small addition of piperine. The authors then filtered and stored the compounds so that they could later be analyzed.

The authors used previously collected extracts from leaves and analyzed them using gas chromatography mass spectroscopy (GC-MS), which determines chemicals within an extract based off particular peaks given off by those chemicals. They were able to find that 40% of all Piper species contained the same types of chemicals.

In order to conduct the experiment, patches needed to be sampled and the average number of Piper per patch needed to be calculated. The researchers created transects located between 50 and 100 m from the trail and a large space separated each transect (more than 250 m). Then in a different transect, 81 plots of 20m in diameter were created. The standardization of the plots is important for an accurate calculation of the average number of Piper per plot.

Piper had different sizes, so for a better standardization, only Piper that are more than 1cm in diameter were selected for calculating the mean number of Piper per plot. The researchers than identified the species of each Piper. A total of 2,035 individuals were counted. The mean per plot was calculated by dividing the number of individuals by the total number of plots. (2,035/ 81 = 25.123) The number of species per plot was calculated by determining the mean of the average of species per plot. The average was found to be 5.2 species per plot.

In the process of his research, Pianka came up with a universal formula for calculating the overlap of pairs of niches.

Here, in situ means "in place." Measurements were made of the actual footprint contours left by the individuals striding along the trackway.

For each test-pit, the authors analyzed the soil profile.This established how the different areas were related geologically and provided information about how to proceed with the analysis of hominin and wildlife print present.

The degree of fossilization of the footprints varied in different locations. In order to study the shape and structure of a footprint, soil had to be carefully removed from the depression.

The trackway footprints are extremely fragile. Clearing and preserving them had to be done with the utmost care. Unlike fossil bones that are solid, the footprints are only empty spaces. Depending on the substrate, an incorrect move could cause the soil to crumble and the footprint to be destroyed. The author had to do a careful analysis of the trackway soils to be able to know how to best preserve and analyze each footprint.

Site G is the code used by Leakey to identify the trackway her group discovered. The three individuals leaving the footprints are referred to as G-1, G-2, and G-2. Keeping with this code system, the authors identified the new trackway area as Site S. The footprints are attributed to individuals S-1 leaving foot prints numbered S-1-1 through S-1-7 in location L8. Individual S-1 also traveled through locations M9 and TP2.

It is only by embracing Newton's ideas that physicists such as Einstein were able to consider exceptions and, in doing so, build on them.

Although we don't often think about it, scientists improve on each others' ideas all the time and revise theories in light of new evidence generated by new technologies and methodologies.

Einstein introduces the need for a concept of general relativity, which describes motion with respect to any two coordinate systems, not just inertial coordinate systems.

Einstein's question points out that coordinate systems are not a product of nature, but a way of understanding nature. It is because of this that physicists are able to revise the tools and methodologies they use in light of new evidence.

After determining that neither the SM tag nor Fab would disrupt cellular processes, the authors wanted to determine how soon Fab would mark KDM5B. To do this, the authors repeated their first trial but imaged at 6 hours instead of 24.

Preliminary testing was conducted twenty-four hours after the plasmid was transiently transferred to ensure that the methods of coupling the FLAG SM tag (that can be labeled with the anti-fluorescein antibody, anti-FLAG Fab) and the MS2 stem-loop (that can be identified with a MS2 coat protein) would not inhibit transcription, translation, or the movement of SM-KDM5B mRNAs throughout the cell.

The results are highlighted in Fig. 1C with the SM-KDM5B protein in green and the MCP-labeled mRNAs in red.

JF646 produces a brighter fluorescence than other proteins.

MS2 is a protein from the coat of bacteriophages. It has a high affinity for "stem-loop" structures, which are hairpin-like shapes formed by DNA. The authors took advantage of this high affinity by using labeled MS2 to detect certain sequences.

Scientists are always working to improve existing techniques to be more precise and versatile.

The final map, 3H, uses a different color coding key, and shows the grid points that are different between the past, present, and future scenarios. Grid points that differ between the 4700 B.P. scenario and the present scenario are marked with blue circles. The filled dots all indicate differences in the biomes between the present and certain future scenarios. Yellow dots are grid points that are only different in the 4.0°C scenario, RCP8.5. The orange dots are different in both the RCP4.5 and RCP8.8 scenarios, and the red dots are different in the RCP2.6, RCP4.5, and RCP7.5 scenarios.

The third map, 3C, was generated by putting the current climate data into the BIOME4 model, running in "forward" mode. This does differ from the first map, based on pollen core samples. The biggest difference is that the first map shows warm mixed forest whereas the third map shows Temperate Deciduous Forest; this is likely caused by pine pollen being over-represented in the pollen core data.

The second map, 3B, comes from putting the 4700 B.P. pollen core data (covering the years 4650–4750 B.P.) into the BIOME4 model. This time slice was picked to represent the past Holocene because it had the highest variation from the present distribution - this is the bar that extends above the 99th percentile line in Figure 2.

These lines, marked below in purple, provide information about the distribution of the black bars, which show the proportional biome change in past Holocene compared to the current period. Half of the values are at or below the lowest line, whereas all but 1% of the past values fall at or below the top line.

Here's the figure with the lines more prominently marked:

To calculate future values, BIOME4 was run in its normal mode—using climate inputs to generate biome predictions. The climate inputs come from the models and time-series data found in the Coupled Model Intercomparison Project (CMIP5), which "promotes a standard set of model simulations." This way, research is done using the same underlying models, which allows for better comparisons across projects. Researchers update these models approximately every 5 years, incorporating new information.

The researchers use two main methods to generate the information about the Mediterranean Basin in this paper—for the past, the model uses data from pollen cores to generate information about the climate and ecosystems found over the past 10,000 years. For the future, the model uses different CO₂ emissions pathways to generate projections about the ecosystems and climate.

The error bars for the Holocene and 1901–2009 data points are based on standard deviations. This is a different method from what the researchers used to calculate the future predicted values.

Authors found that this species of ants are more efficient at being predators.

The ant species C. floridanus death rate is at 56.3% during this experiment. They have forced some or all instar caterpillars to leave their habitat.

The instar caterpillars were vulnerable to be eaten, or killed by other predators.

Each species was investigated to determine and compare each tree group.

The experiment was conducted at Biscayne National Park and was split into three parts. The first part was using pitfall traps to collect ants at Biscayne Park near the host plants. Secondly, the host plants with known caterpillars were protected with various combinations of cages and tanglefoots. Concluding the author was keeping track of how long it take for certain species of ants to first find the caterpillars. Authors assumed that the ants would interact with caterpillars more frequently than other predators.

All of the ants found in pitfall traps that were placed all around the canopy in the keys. 11 out of 25 species were found to be exotic and 735 out of 1418 of the total amount of ants were exotic.

After the experiment was finished, they have found that the Pseudomyrmex gracilis and C. floridanus interacted more aggressively toward the subject compared to another species.

A learned association between a stimulus and negative outcomes, in this case represented by pictures associated with an electric shock.

Learn more about fear conditioning (animal study): https://www.youtube.com/watch?v=ozkpK-nhz04

Reinforcement that allows the association of a symbol with a positive outcome, in this case money. After trials, the objective was to teach participants to respond to the animal pictures associated with a reward.

This task establishes an association between a neutral stimulus and a negative outcome with the hope that the repetitive task will teach participants to avoid the stimulus.

In this study, the negative stimulus was an electric shock over four stages of learning.

The determination of the correlation coefficient.

In this study, the authors used both Pearson (r) and Spearman (rho) correlations. Pearson correlation is used when there is an assumed normal distribution (distribution that shows symmetry around the mean), while Spearman is used when it is assumed the distribution will not be normal.

Learn more about coefficient determination:

http://blog.uwgb.edu/bansalg/statistics-data-analytics/linear-regression/what-does-coefficient-of-determination-explain-in-terms-of-variation/

The eight blocks (96 trials) where participants were tested and learned from the feedback that they received after their response to the presentation of the pictures.

A urine screen is performed to rule out medical conditions and/or drug use. In this case, the authors used it to determine if the volunteers were appropriate for their study and to see if they tested positive for drug use for the different types of drugs involved.

A test that makes it easier to analyze data that aren’t normally distributed. The authors performed this analysis to reduce skewness (effects due to asymmetry in a probability distribution) and stabilize variance from the questionnaires and demographic responses.

To learn more about transformation: http://fmwww.bc.edu/repec/bocode/t/transint.html

In this experiment, skin from a human was grafted onto mice with depressed immune systems. This means a piece of human skin was transplanted onto mice whose immune systems could not protect them from diseases.

In this study, the cells from the neural crest of mice that carried the specific mutation Kitl were treated with endothelin 3.

The study described here consisted of treating quail (type of bird) neural crest cultures with Edn3. This study was in vitro, meaning the embryonic neural crest was removed and treated outside of the organism. This study showed the role that Edn3 plays in development as stated in the paper.

Exact numbers are not available for the mass of plastic waste entering the ocean, so the authors used a series of estimates in their calculations.

For the reconstructions of the past Holocene ecosystems based on the pollen cores, the initial model outputs are at a lower resolution than the 0.5° by 0.5° resolution of the future projected biomes. The resolution of the reconstructions was a grid of 2° latitude and 4° longitude.

To allow more direct comparisons between past and future, the researchers interpolated the reconstruction data to fill in a 0.5° by 0.5° grid. However, this may miss some features at scales smaller than the original gird.

How the authors obtained AGB and tree biodiversity, evaluated trade off among tree diversity and aboveground biomass

Table 1 shows the different types of plots used and how they differed in size, area covered, areas inventories, shape, and number of days used.

Fundamental paper which shows how plants will uptake more CO2 when it is present in higher and higher quantities.

At 600 parts CO2 per million parts air, plant carbon uptake is limited due to low levels of CO2 present in atmosphere.

The solid bars show the actual data gathered. They measure the different rates of carbon exchange.

Figures displays projections for the effect of three climate-induced environmental changes on ecosystem processes (NEE, Reco, GEE).

As CO2 levels and temperature increase, the ecosystem will uptake more and more CO2, and will retain more carbon.

The figures show what happened to the actual data vs the data from the model. They are very close and show that the model is reliable, in other words, the data is precise.

Remember, this is a simulation! It was done on the DAYCENT simulation model.

These charts show the different magnitudes, at differing scales of CO2, temperature, and rainfall, respectively, projected out to the year 2076.

By comparing a modeled map using every data point available and comparing that map to one that excludes outliers exceeding 3 standard deviations, the authors realized they were actually excluding some ocean temperature features that occurred. Thus by changing the data cut-off to 6 standard deviations it was possible to more accurately model ocean conditions.

By broadening the cutoff criteria the authors of this study were able to distinguish real-life changes in ocean temperatures. Their previous data cutoff of 3 standard deviations (which are a measure of variance around a mean), was too strict and deleting real data points.

Broadening the pool of what was considered "good" or "realistic" data to 6 standard deviations was necessary in order to properly document temperatures that actually occurred in the oceans.

Biological factors that drive the decomposition of carbon include the consumption and respiration from microbes, vertebrates, and in-vertebrates.

Physical factors include temperature, flow of water, and sediment.

The alignment of the reach-litterbag scale rates show that the bags represent the actual rates, and that the bags were not simply filtered by increased water flow or torn apart because of temperature.

The litterbag technique can be used to measure the density lost due to composition by microbes and fungi.

A known type and amount of litter is stuffed in the bag and is left for a certain amount of time in the stream. The bag is then later taken out and weighed while wet. The bag is allowed to dry and then the dry weight is taken.

Benthic fine and coarse particulate organic carbon can be used as a measurement when observing loss of carbon caused by detrivores.

In this experiment, leaf litter from different types of trees were used. The results are in Figure 3.

The structure and chemistry of leaves are different and this determines how fast they may decompose.

For example, tulip poplar decomposes the quickest and has a low carbon to nitrogen ratio. Comparatively, oak decomposes slowly and has a higher carbon to nitrogen ratio.

It is often necessary to use a simpler model version of a larger system because it is easier to observe.

It is difficult to observe the effects on a whole stream, but the litter bags are observed and represent the carbon loss in streams.

Litter in the litter bags have a higher concentration of carbon than the stream, which has carbon more spread out. The litter will have higher loss of carbon but generally represents reach-scales.

Two experiments were done separately to test the effects of nitrogen to phosphorus ratios on streams.

Pre-treatment of streams include recording the levels of nutrients and typical conditions throughout a year.

The first experiment had two watersheds with one being the control and the other having an addition of nitrogen and phosphorus to match a ratio that was decided by the scientists.

The second experiment was done on five streams with different combinations of the nitrogen and phosphorus ratio.

Although this experiment focuses on microbial and fungal effects on streams, humans sometimes also have an impact.

These effects may include the leaking of septic tanks into bodies of water that increase human waste and phosphorus levels. Fishing and polluting also affect stream ecosystems.

The experiments were conducted on streams in the Appalachian Mountains of North Carolina.

Because of the elevation, there are little to no fish influencing the data. Mainly microbes and fungi influence the nutrient levels of the streams.