Galaxy Zoo CANDELS: Morphologies of 50,000 z ≲ 4 Galaxies Brooke D. Simmons 1,2, Chris Lintott 1, Karen Masters 3, Kyle Willett 4, Jeyhan S. Kartaltepe 5, Henry C. Ferguson 6, Sandra M. Faber 7, Boris Häußler 8, and the Galaxy Zoo and CANDELS science teams 1 University of Oxford, 2 UCSD, 3 University of Portsmouth, 4 University of Minnesota, 5 RIT, 6 STScI, 7 UCO/Lick, 8 ESO Over 95,000 Galaxy Zoo volunteers contributed over 2,000,000 independent visual classifications of 49,555 IJH images (selected to have H < 25.5) from three CANDELS fields: GOODS-South, COSMOS, and UDS. The classifications have been analyzed and a public data release is imminent (Simmons et al., submitted). More Classification of Detailed Features Overall Demographics Less

Is the galaxy simply smooth and rounded, with no sign of a disk? Smooth Features Star or or disk artifact Smooth

How rounded is it? Does the galaxy have a mostly clumpy appearance? Completely In Cigar Yes No Classification flow A majority of featured round between shaped diagram: for each galaxies have a

classification task clumpy appearance, Features we show the and most clumpy How many clumps are there? Could this be a disk viewed edge-on? 1234More Can't Yes No proportions of galaxies are than 4 tell plurality answers asymmetric with a

for the galaxy mix of clump Artifact population in counts. Among paths leading to spiral galaxies, 2 Do the clumps appear in a straight Is there a sign of a bar feature subsequent tasks. arms are typical. line, a chain, or a cluser? through the center of the galaxy? Cluster Spiral Straight Chain Bar No bar Each task only These charts include Line includes galaxies galaxies at all Clumpy where the question redshifts and is relevant (example: luminosities; sub- the Clumpy task only samples selected via

includes galaxies different criteria (e.g., Edge-on Is there one clump which is clearly Does the galaxy have a Is there any sign of a spiral arm considered “featured” luminosity, mass) will brighter than the others? bulge at its center? pattern? No Yes Yes No Spiral No spiral in the first task). likely have very different overall demographics. Spiral

Is the brightest clump central How tightly wound do the to the galaxy? spiral arms appear? Yes No Tight Medium Loose Merger

Does the galaxy appear Weighted classifier inputs are combined into vote fractions f for each galaxy and feature, including symmetrical? How many spiral arms are there? Yes No 1234More Can't tell mergers, bars, and clumpiness. Each row of inverted IJH images shows examples in a specific than 4 morphological category. (0.ʺ06 / pixel). Histograms show distribution of vote fractions for each response in a given row, considering galaxies with appropriate classifications in parent tasks.

Comparison with professional visual classifications Do the clumps appear to be Featureless disk galaxies at z > 1 embedded within a larger object? How prominent is the central bulge, compared Galaxy Zoo (typically ≳ Yes No with the rest of the galaxy? 40 classifications per No bulge Obvious Dominant galaxy) and CANDELS team classifications (typically 3 per galaxy; Kartaltepe et al. 2015, K15) correlate strongly

st over a range of features. 1 Tier Question Is the galaxy currently merging or is there any sign of tidal debris? nd Merging Tidal Both Neither The agreement is 2 Tier Question debris 3rd Tier Question remarkably good 4th Tier Question despite the two 5th Tier Question classification methods and questions typically being quite different. In End some cases, using a combination of indices Classification tree: a single classification in Galaxy Zoo CANDELS consists of 17 separate tasks, from both catalogs may each answering a question about the subject. Morphological measurements include details of be most effective in clump structures, bar features, spiral arm counts, bulge strength in disks, and merger/tidal features. selecting a pure sample of galaxies (e.g., Comparison of bulge-to-total ratios (Häußler et al., in preparation) of “smooth” vs “featured” galaxies at mergers). 1 < z < 3 to those at z ~ 0 (Lackner & Gunn 2012) reveals an additional population of z > 1 disk- This work is made possible by the participation of more than 95,000 volunteers in the Galaxy Zoo project. The contributions of the dominated galaxies lacking any features typically associated with disks, compared to local average average more than 40,000 of those who registered a username with Galaxy Zoo are individually acknowledged at http://authors.galaxyzoo.org/. galaxies with the same luminosities relative to the evolving optical galaxy luminosity function and binned binned by GZ by K15 sampled at redshifts corresponding to the same physical resolution in kpc in SDSS images (z ≈ 0.04).