For full functionality of this site it is necessary to enable JavaScript. Here are the instructions how to enable JavaScript in your web browser.

Click for links to photos, slides and videos now available from the 2018 FaceBase Annual Meeting.



Why make FishFace?
Abbreviations Source
How we made FishFace

Why make FishFace?

How do the elements of the craniofacial skeleton arise, grow, and reshape? Answers to this question are coming from both molecular-genetic and cell-biological approaches, which rely, first of all, on precise description of the developmental events and processes that comprise skeletogenesis. Zebrafish, with a sophisticated knowledge of its genetics and genomics, with favorable attributes for phenotypic analyses of development, and with patterns of development conserved among all vertebrates, provides a powerful animal model for learning about craniofacial development. In particular, with current transgenic approaches one can examine craniofacial skeletal elements in exquisite cellular detail during an extended period of development within living, intact embryos and larvae – an investigative method unsurpassed in accuracy and sensitivity. We constructed this developmental atlas of the craniofacial skeleton, FishFace, to serve as a guide for such study.

We hope that the FishFace Atlas will be particularly useful in comparative and mutational analyses where there is interest in understanding the cellular basis of early skeletogenesis. The heart of the FishFace Atlas uses high magnification (generally a 40x objective) confocal image stacks showing transgenically-labelled chondrocytes or osteoblasts, along with mineralized bone matrix, which is visualized by vital staining with Alizarin red. We present these stacks in sequences that follow particular individual cartilages and bones of the first two pharyngeal arches as they develop during embryonic and larval stages (see: How to navigate the Atlas pages). To do so, we build on the foundation set out in the gold standard reference for describing comprehensively skeletal elements in the zebrafish craniofacial complex, Cubbage and Mabee (1996), which used fixed preparations stained for cartilage and bone through adult stages.

The FishFace Atlas element development section adds considerable detail to arch one and two early development, particularly at the cellular level, but also in description of element growth and shaping. Other sections of the FishFace Atlas, at lower magnification, provide anatomical context for the element development section, including an interactive tool made by optical projection tomography (OPT) for learning the anatomy of the entire larval skull. Hence, the FishFace Atlas provides the community with an interactive resource with which the user can understand not only the cellular details, but also complex 3D anatomical relationships, of developing elements in the craniofacial skeleton of the zebrafish.

Abbreviations Source

The list of abbreviations for anatomical terms, which is based largely upon terminology of Cubbage and Mabee (1996). can be found here. Also included are the ZFIN anatomical definitions for these terms, as well as links to images in FishFace that demonstrate each particular anatomical term.

How we made FishFace

Transgenic animals
  1. zc81Tg labels chondrocytes. This transgenic line was created using Gateway technology and includes a portion of the foxp2 promoter driving EGFP expression (Tg(foxp2.A:EGFP)) (Bonkowsky et al., 2008). However, the published fish line had two major domains of expression: 1) brain, and 2) cartilage. The brain expression was expected from the known foxp2 expression patterns, but the cartilage expression was unexpected, as foxp2 is not expressed in this tissue.We were able to isolate fish that had only brain or cartilage expression, suggesting that two insertions of the Gateway construct were responsible for the two initial transgene expression domains. Thus, the original line with brain expression is Tg(foxp2.A:EGFP)zc42, and the line in which we isolated the cartilage expression is Tg(foxp2.A:EGFP)zc81. Since we could not replicate the cartilage expression by injecting the Tg(foxp2.A:EGFP) construct, and there was only one initial founder with the cartilage expression, we assume that the construct is expressed in cartilage due to position-dependent effects, similar to an enhancer trap. Our experiments to identify the genomic locus of the Tg(foxp2.A:EGFP) insertion by inverse PCR and genetic linkage suggest that the insertion site is linked to the known chondrocyte differentiation gene sox9a (data not shown). Therefore, the formal name of this line is sox9azc81Tg, but we will refer to it in this Atlas as zc81Tg.
  2. Tg(sp7:EGFP)b1212 labels osteoblasts. This transgenic line was created using BAC transgenesis and includes a large portion of the sp7 promoter driving EGFP expression (DeLaurier et al, 2010).  We refer to it in this Atlas as sp7:EGFP.
  3. Tg(fli1a:EGFP)y1 labels neural crest cells. This transgenic line was created by micro-injecting a construct including a portion of the fli1a promoter driving EGFP expression (Roman et al., 2002) and has been used widely as a marker of neural crest cells and their derivatives. We refer to it in this Atlas as fli1a:EGFP.
Alizarin red staining

For confocal imaging of mineralized matrix in live specimens, larvae were maintained in Embryo Medium supplemented with 0.005% Alizarin red and 0.01M HEPES at least two hours, and often overnight, prior to confocal imaging (Kimmel et al, 2010). For OPT imaging of mineralized matrix in fixed specimens, larvae were stained as described previously (Eames et al., 2011). Tissues were fixed overnight in 4% PFA, washed for an hour in 1% KOH, bleached in 3% H2O2/0.5% KOH for 40 min. with lids open, washed in 1% KOH, stained overnight in 0.003% Alizarin red in 1% KOH, and de-stained in 1% KOH. After eyes were removed, heads were embedded in agarose, washed twice in methanol, and cleared in benzyl alcohol:benzyl benzoate (1:2).

Confocal microscopy

Imaging of live specimens was conducted using either a Zeiss LSM 5 Pascal confocal or Leica SD6000 spinning disk confocal with Borealis illumination technology. In order to visualize Alizarin red staining with most sensitivity, pinhole and/or detector gain settings were adjusted by hand to levels just below those that showed red fluorescence in surrounding, non-mineralized tissues. Maximum projections were made from stacks of images that demonstrated the entire depth of the element(s) under focus. Movies showing progression through all images of the stack were created in Pascal (Carl Zeiss, Inc.) or Metamorph (Leica Microsystems).

OPT imaging and processing

Images were captured using Bioptonics OPT Scanner 3001M (MRC Technology).  3D reconstructions of raw data were made using NRecon (MRC Technology).  Movies showing progression through virtual sections of the reconstruction were created in QuickTime Player (Apple).  Segmentations of skeletal elements in reconstructions were made using Amira 5.2.2 (Visage Imaging), and movies highlighting specific skeletal elements were created in Amira and QuickTime Player.



The FishFace Atlas provides a reference for the anatomy of the zebrafish craniofacial skeleton. It also illustrates a generalized developmental sequence of the appearance of specific skeletal elements, but the reader should be aware of caveats to such generalizations. First, development is a dynamic, variable process. Our temporal series of images are merely snapshots of this process and thus can not represent the full dynamics of developmental processes. Second, considerable clutch variation exists within a given ‘wild-type’ stock, let alone between stocks in use around the world. We have attempted to normalize any such differences by noting the standard length of each specimen. In addition, we choose for display in the FishFace Atlas those specimens that represent the typical size and mineralization in a clutch at each timepoint. Finally, we are subject to limits of detection, so we try to avoid making declarative statements that a given skeletal element ‘appears’ at a given time. That said, we have refined methods to detect skeletal elements with the most sensitive techniques available. We image specimens with the vital dye Alizarin red while they are alive, and our transgenesis and confocal imaging illustrate skeletogenic cells prior to their secretion of abundant extracellular matrix. As such, we avoid the severe reduction in Alizarin red binding that occurs after even brief periods in fixative and artificial buffers, for these chemicals leech mineral from skeletal tissues in the specimen. In summary, the FishFace Atlas addresses potential pitfalls of a variable developmental process and limitations of skeletal detection to provide the user an understanding of development of the pharyngeal arches and its skeleton in the embryonic and larval zebrafish.


Two things about fish skeletal anatomy deserve mention to assist the user of the FishFace Atlas. First, the nomenclature of skeletal anatomy can vary considerably from decades of various descriptions, and we base our terminology on that employed by Cubbage and Mabee (1996)—please see our Abbreviations listing. In a few cases, the name given for a cartilage and its surrounding perichondral bone is the same, which can be confusing. For example, the ceratohyal can be both a cartilage and the bone that forms in the perichondrium of the middle portion of the ceratohyal cartilage. To help avoid confusion in these cases, we abbreviate the ceratohyal cartilage with ch, and the ceratohyal bone with chb. The basihyal cartilage (bh) and bone (bhb) is the other example among the elements that we illuminate in the FishFace Atlas. Second, a given bony skeletal element may initiate in the perichondrium, but may later extend into the dermis. This has been reported to be a common feature of ossification in various fish (Cubbage and Mabee, 1996), but is unfamiliar to specialists in mammalian skeletogenesis. For example, the hyomandibula initiates in the perichondrium of the dorsal hyosymplectic around 5 dpf, but by 14 dpf, a clear extension of the hyomandibula from the perichondrium is seen to run in the dermis laterally. Given this feature of fish skeletogenesis, we classify a bone on the Abbreviations page as either chondral or dermal depending upon the location of its first ossification.

Other sites of potential interest to FishFace Atlas viewers

Zebrafish Information Network (ZFIN) —provides useful definitions of anatomical terms and integrates mutant zebrafish data to identify genes that may affect specific anatomical structures; also contains other Atlases and embryonic staging series site supported by University of Oregon. 

Zebrafish Atlas — histological slides encompassing all zebrafish life stages by Penn State’s Jake Gittlen Cancer Research Foundation

FishNet — annotated 3D renderings from OPT data of embryonic and larval zebrafish by Monash University’s Biological Sciences

Interactive Atlas of Zebrafish Vascular Anatomy — fluorescent angiograms from 1-7dpf zebrafish by NIH’s NICHD

Interactive 3D Mouse Limb Anatomy Atlas — annotated 3D renderings from OPT data of immunofluorescently-stained E14.5 mouse limbs by NIMR(MRC)

Quail Developmental Atlas — annotated 3D renderings from MRI data of quail embryos from 5 to 10 days by Caltech’s Biological Imaging Center


1) Bonkowsky et al. (2008) Domain-specific regulation of foxP2 CNS expression by lef1. BMC Dev Bio 8: 103

2) Cubbage and Mabee (1996) Development of the cranium and paired fins in the zebrafish, Danio rerio (Ostariophysi, Cyprinidae). J Morph 229: 121-60

3) DeLaurier et al. (2010) Zebrafish sp7:EGFP: a transgenic for studying otic vesicle formation, skeletogenesis, and bone regeneration. Genesis 48(8): 505-11

4) Eames et al. (2011) Mutations in fam20b and xylosyltransferase1 reveal that cartilage matrix controls timing of endochondral ossification through inhibition of chondrocyte maturation. PLoS Genetics 7(8): e1002246

5) Kimmel et al. (2010) Modes of developmental outgrowth and shaping of a craniofacial bone in zebrafish. PLoS One 5(3): e9475

6) Roman et al. (2002) Disruption of acvrl1 increases endothelial cell number in zebrafish cranial vessels. Development 129(12): 3009-19