DATA FROM: “Larger but not louder: bigger honey bee colonies have quieter combs” README FILE FOR OPEN DATA We have compiled the data collected for the manuscript titled “Larger but not louder: bigger honey bee colonies have quieter combs.” If other researchers would like to use these data, they are free to do so (see license information below). We only ask that they reference the publication, and the data. Suggested citation for the publication: Smith, ML, & Chen, P (2017) Larger but not louder: bigger honey bee colonies have quieter combs. Behav Ecol Sociobiol 71:169. https://doi.org/10.1007/s00265-017-2399-9 Suggested citation for the data: Smith, M.L., & Chen, P. (2017) Data from: Larger but not louder: bigger honey bee colonies have quieter combs. Cornell University Library eCommons Repository. https://doi.org/10.7298/X48050Q2 If researchers would like to collaborate using these data for further analyses, we recommend they get in touch with the corresponding author, Michael L. Smith, mls453@cornell.edu. There are three folders in the open data, each folder contains data from the measurements of comb vibrations. One folder has the raw accelerometer data (.txt), one folder has the .wav files which were created from the raw accelerometer data, and one folder has the data for the trial wherein we added dead bees to the comb (both .txt and .wav files in that folder). The last two files are .csv spreadsheets containing the data that we used to compare vibrations in the comb. These spreadsheets contain the metrics from the .wav files, calculated in Raven Pro, as well as the measurements of colony size. The file “SmithEtAl_BehavEcolSociobiol_2017_WavConvert.m” is the MATLAB script that we used for converting .txt files to .wav files. This script was used in MATLAB version R2015b. DATA LICENSE: This data is being shared under a CC-By 4.0 International license (https://creativecommons.org/licenses/by/4.0/). Data may be used, shared and adapted freely, but proper attribution to the original authors must be given. MANUSCRIPT INFO: TITLE: Larger but not louder: bigger honey bee colonies have quieter combs AUTHOR INFO: Michael L Smith (1)* Po-Cheng Chen (2) Author Institutions and Contact Information: (1) Department of Neurobiology and Behavior, Cornell University, Ithaca NY, USA MLS453@Cornell.edu +1 281-686-4717 ORCID: 0000-0002-3454-962X (2) School of Electrical and Computer Engineering, Cornell University, Ithaca NY, USA PC445@Cornell.edu * Corresponding Author ABSTRACT: Communication is impossible if the sender’s signal cannot overcome background noise to reach the receiver. This obstacle is present in all communication modalities, forcing organisms to develop diverse mechanisms to overcome noise. Honey bees will modify combs to improve signal efficiency of substrate-borne vibrations, but it is unknown whether, and if so, how, bees compensate for the largest potential source of noise: the bees themselves. The number of bees in a colony changes markedly throughout the year, but the size of the nest cavity does not, forcing workers into high densities on the combs. How, then, do bees communicate via substrate-borne vibrations on combs that are covered in bees? We used accelerometers to measure comb vibrations, while varying the number of workers on the comb. Surprisingly, comb vibrations decreased with increased worker number. Furthermore, inserting freshly killed bees to the comb demonstrated that it is not simply the bees’ collective mass that damps vibrations, but is probably their behavior. We propose that their posture damps vibrations, with each bee linking up to six neighboring cells with her legs. This collective damping reduces background noise, and improves the landscape for communication. These results demonstrate how living systems, including superorganisms, can overcome physical obstacles with curiously simple and elegant solutions. KEYWORDS: Substrate-borne vibration, signal propagation, honeycomb, superorganism, social insects, colony size FUNDING: MLS is supported by the National Science Foundation Graduate Research Fellowship Program (DGE-1144153). This research was funded with a National Science Foundation Doctoral Dissertation Improvement Grant (1600775), an Andrew W. Mellon research grant, and a Centennial Pollinator Fellowship from the Garden Club of America (to MLS). OPEN DATA FOLDERS AND FILES: FOLDER CONTAINING ACCELEROMETER DATA (.txt files) The.txt files are the raw data from the 3-axis accelerometers, measured in units of gravity (g). The headings x(g), y(g), and z(g), denote the three axes, X, Y, and Z. The name for each file gives information about the date the data was collected, and the experimental trial it was part of. [Date]_[Experiment]_[Trial] Date is given in the format DDMonthYYYY, for example: 4July2016 Experiment is a description of the data collection (see methods in manuscript). The baseline measurement is before bees were added to the observation hive, so there are zero bees on the comb. Half Colony denotes when we’ve added half the colony of bees to the observation hive, Whole Colony denotes that we’ve added all the bees to the observation hive. If a TIME is given, eg 10am, then that denotes the time during the day that we collected the data, for the trials in which we measured comb vibrations throughout the day (see methods in manuscript). Trial is for when we took multiple measurements, for example, taking more than one baseline measurement. So, for example, this file: 8July2016_HalfColony_2.txt The comb vibrations were measured on 8 July 2016, with half the colony in the observation hive, and this is the second of two measurements we took at this colony size. NOTE: 1 August 2015 is a whole day of baseline measurements, so the Experiment shows time of day, but each measurement is for comb vibrations with no bees in the observation hive. FOLDER CONTAINING .WAV FORMAT OF THE .TXT FILES The raw data from the accelerometers (.txt format) were converted to .wav format in MATLAB (see methods in manuscript). The naming system for these files is the same as for the raw data. FOLDER CONTAINING DEAD BEES DATA (.TXT AND .WAV FILES) Dead bees were added to a focal comb without any live bees on it (see methods in manuscript). The .txt files are the raw output from the accelerometer, and the .wav files are the reformatted .txt files to .wav format. The naming system for these files is: [DATE]_[Trial]_[Number of dead bees] There is only one date for the dead bee trial, 14 December 2015. The files only list Dec14 for all the files. Trial goes from 1 to 20, denoting the order that the comb measurements were made. “Number of dead bees” denotes the number of dead bees that are on the comb. For example: Dec14__9_1474bees.txt This is raw output from the accelerometer. The sample was taken on 14 December 2015, it was the 9th recording from the comb, and there were 1474 dead bees on the comb. Dec14__20_0bees.wav This is a .wav file, which was converted from the raw .txt format. This sample was taken on 14 December 2015, it was the 20th recording from the comb, and there were 0 dead bees on the comb. RAVEN PRO SOFTWARE CITATION: Bioacoustics Research Program. (2011). Raven Pro: Interactive Sound Analysis Software (Version 1.4) [Computer software]. Ithaca, NY: The Cornell Lab of Ornithology. Available from http://www.birds.cornell.edu/raven. SPREADSHEET CONTAINING DATA ON COMB VIBRATIONS FROM RAVEN PRO, AND COLONY SIZE (.csv) Spreadsheet containing values from the .wav file analysis in Raven Pro, and from the colony size measurements (see methods in manuscript). The file has the following headers: Date The date on which the data were collected, labeled DD-Month-YYYY ColonyNum The Colony Number that the data pertain to. Cnull is for the whole day baseline measurement, wherein no bees were on the comb, and vibrations were measured throughout the day. Trial Trial denotes the measurement from the experiment. For example, for the 24hr experiment, comb vibrations were measured throughout the day. 2pm, therefore, denotes the trial at 2pm, within the 24hr experiment. Baseline denotes the baseline measurements, where no bees were present in the observation hive. Half Colony denotes half the colony is in the observation hive, Whole Colony denotes the whole colony has been installed into the observation hive. If there is a number after the descriptor (Baseline2, for example), it denotes that multiple measurements were taken at that level. Hour The time of day that the measurement was taken, on a 24hr clock. Experiment The experiment that the data pertain to. Whole Day Baseline indicates the day during which baseline measurements were taken all day. HalfColony/WholeColony indicates experiments where we installed half a colony, measured comb vibrations, and then added the rest of the colony, and measured comb vibrations again. Agg_Entropy Aggregate entropy, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. Avg_Entropy Average entropy, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. Avg_Power Average Power, as calculated in Raven Pro from the .wav files. Units for this measurement are in dB. Max_Power Maximum Power, as calculated in Raven Pro from the .wav files. Units for this measurement are dB. Energy Energy, as calculated in Raven Pro from the .wav files. Units for this measurement are dB. Center_Freq Center Frequency, as calculated in Raven Pro from the .wav files. Units for this measurement are Hz. Max_Freq Maximum Frequency, as calculated in Raven Pro from the .wav files. Units for this measurement are Hz. Max_Amp Maximum Amplitude, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. RMS_Amp Root Mean Squared Amplitude, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. ColonySizeGrid The number of bees in the whole observation hive, as measured using the grid technique. FrameSideWacc The number of bees on the focal comb, where the accelerometer was placed. This is for the side of the comb that has the accelerometer. Bees were only counted if they were on the comb itself (not on the glass of the observation hive). NA in this column denotes “Not Available” because the data were not collected. This is because the number of bees was already measured for a previous measurement, so we did not want to double count the number of bees in the colony for a similar colony size. FrameSideWOacc The number of bees on the focal comb, where the accelerometer was placed. This is for the side of the comb that does not have the accelerometer. Bees were only counted if they were on the comb itself (not on the glass of the observation hive). NA in this column denotes “Not Available” because the data were not collected. This is because the number of bees was already measured for a previous measurement, so we did not want to double count the number of bees in the colony for a similar colony size. AccFrameTotal The total number of bees on the focal comb where the accelerometer was placed. This is the measurement that we used for the x-axis in the plots presented in the manuscript. It is also the metric that was used for statistical analyses. NA in this column denotes “Not Available” because the data were not collected. This is because the number of bees was already measured for a previous measurement, so we did not want to double count the number of bees in the colony for a similar colony size. ColonySizeWt The size of the colony, as measured by the weight of the bees. These values are only valid for the experimental trials where half the colony was added to the observation hive, and then the whole colony was added to the observation hive. “NA” in this column denotes “Not Applicable” because the weight of the bees in the colony is unknown. SPREADSHEET WITH DEAD BEES DATA (.csv) Date The date on which the data were collected, labeled DD-Month-YYYY (Day-Month-Year) ColonyID The Colony ID for this case is the same, deadbees. There is no colony in the observation hive, only dead bees. Trial Trial denotes the trial of the dead bees experiment, in order from 1 to 20. DeadBeesOnFrame The number of dead bees in the comb of the frame with the accelerometer. Note that the number of dead bees increases in even numbers up to 1600, and then decreases. That is because we first loaded the comb with dead bees, measured comb vibrations as we added dead bees to the comb, and then removed the dead bees, and took additional measurements as varying numbers of dead bees had been removed. Agg_Entropy Aggregate entropy, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. Avg_Entropy Average entropy, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. Avg_Power Average Power, as calculated in Raven Pro from the .wav files. Units for this measurement are in dB. Max_Power Maximum Power, as calculated in Raven Pro from the .wav files. Units for this measurement are dB. Energy Energy, as calculated in Raven Pro from the .wav files. Units for this measurement are dB. Center_Freq Center Frequency, as calculated in Raven Pro from the .wav files. Units for this measurement are Hz. Max_Freq Maximum Frequency, as calculated in Raven Pro from the .wav files. Units for this measurement are Hz. Max_Amp Maximum Amplitude, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless. RMS_Amp Root Mean Squared Amplitude, as calculated in Raven Pro from the .wav files. Units for this measurement are unitless.