Size changes of brain and brain regions along altitudinal gradients provide insight into the trade-off between energetic expenditure and cognitive capacity. We investigated the brain size variations of the Asiatic Toad (Bufo gargarizans) across altitudes from 700m to 3200m. A total of 325 individuals from 11 sites and two transects were sampled. To reduce confounding factors, all sampling sites within each transect were within a maximum distance of 85km and minimum altitudinal difference close to 2000 m. Brains were dissected and five regions were both measured directly and with 3D CT scan. There is a significant negative correlation between the relative whole brain volume (to snout-vent length) and altitude. Furthermore, the relative volumes (to whole brain volume) of optic tectum and cerebellum also decrease along the altitudinal gradients, except the telencephalon, which increases along the altitudinal gradients. Therefore, our results are mostly consistent with the expensive tissue hypothesis and the functional constraint hypothesis. We suggest that most current hypotheses are not mutually exclusive and data supporting one hypothesis are often partially consistent with others. More researches on mechanisms are needed to explain the brain size evolution in realistic adaptation events.

A total of 325 individuals from 11 sites were collected during breeding seasons of 2018 and 2019. We sampled two transects. Transect 1 is located near the Wolong National Natural Reserve and we sampled five sites within a distance of approximately 60 km. Transect 2 is located near the Mt. Gongga and we sampled six sites within a distance of approximately 84 km. In general, we tried to sample sites that were altitudinally 300 m to 500 m apart along a valley when possible. To minimize potential geographic confounding factors, sites within the same transect were selected within the shortest distance possible. To avoid autocorrelation, the two transects were located at two different mountain ranges with the closest points more than 140 km apart. We sampled breeding populations to control potential effects of age and seasonality on brain morphology. The breeding season of this species is at the very beginning of the active season; for populations at low altitudes, it lasts from the end of December to early February, while for high altitude populations the season lasts from April to May. Thus, sampling times of different sites were in different months as the phenology was delayed along altitudinal gradients. All animal procedures were carried out in accordance with the approved protocols from the Animal Care and Use Committee at the Chengdu Institute of Biology, Chinese Academy of Sciences (Permit number: 20180820).

All individuals were euthanized with 0.25% MS-222 solution. Specimens were fixated and stored in neutralized 10% formalin. After three months in storage, specimens were photographed in both dorsal and ventral views with an Olympus camera (Em5mark2). The coronal plane of all specimens was ensured to be parallel to the sensor of the camera and the focal length was fixed. A ruler was placed in all photographs as a reference. Measurements were taken from the photos using ImageJ (v1.53d). The snout-vent length (SVL) was measured to the nearest 0.001 mm, and the measurement was repeated three times for each individual.

Brains were dissected out and the cranial nerves, pineal organ, and meninx were removed. The pituitary glands were also removed but the pituitary infundibulum was preserved. Similarly, photographs of these brains were taken from dorsal, lateral, and ventral views. The coronal plane and sagittal plane were parallel to the camera sensor, and a ruler was placed in all photographs as a reference.

Length (L), width (W), and height (H) of brain and brain regions, including olfactory bulb (OLF), telencephalon (TEL), optic tectum (TEC), cerebellum (CER), and pituitary infundibulum (PIT), were distinguished and measured to the nearest 0.001 mm from the photographs. Each of the traits was measured three times. All the measurements were conducted by the same investigator (ZY) to eliminate inter-observer variability. The volumes of brain and brain regions were calculated using the ellipsoid model as V= (L×H×W) π/ (6×1.43). For OLF, TEL, and TEC, which had two symmetrical structures, only the right-side structures were measured and the volume of the ellipsoid model was multiplied by two. Several brains or brain regions with damages were measured only on one side, so the total volumes were extrapolated assuming the brain is a symmetrical structure. Overall, a total of 268 specimens were measured for whole brain volume (V_LM), 264 for OLF, 264 for TEL, 267 for TEC, 266 for CER, and 235 for PIT.

Eight brain specimens per sampling site (4 females, 4 males) were randomly selected for CT scans to estimate brain volume using 3D modeling (V_3D). Site 2.5 had only six samples and was excluded from this set of analysis. The brains were washed in tap water for 24h and then soaked in I₂KI solution (3.00% w/v) for 45h. A micro-CT machine (PerkinElmer, Quantum GX) was used to obtain the DICOM data at 70 kV, 88 μA.

A fixed threshold was used to obtain the 3D models of brains. The olfactory nerves and other parts that were not included in measurements were removed as in Figure 3b. For brains that were incompletely dyed with iodine, the models were repaired using the symmetrical half of the brain. A total of 80 brain models were constructed and measured. All of the modeling, segmentation, and volume counting was conducted in 3DSlicer (version 4.10.2).