[Bug]: quiver plotting behaves bizarrely with polar plots

Nobody has claimed this yet.

### Bug summary

There is no documentation, as far as I can tell, explaining how quiver should behave with polar plots, and using `pyplot.quiver` with a polar plot results in odd plotting behaviour.

### Code for reproduction

```python
import numpy as np
from matplotlib import pyplot as plt

fig, ax = plt.subplots(subplot_kw={"polar": True})

longitude = np.arange(5, 365, 10)
colatitude = 90 - np.ones_like(longitude) * 60

sample_x = np.ones_like(longitude) # Northward
sample_y = np.ones_like(longitude) # Eastward

q = ax.quiver(np.radians(longitude), colatitude, sample_x, sample_y, angles="xy")
```

### Actual outcome


### Expected outcome

I would expect the code to plot arrows which were vectors pointing all pointing northeast; i.e., I expect `sample_x` to be defined as the northward (poleward) vector and `sample y` to be defined as the eastward (azimuthal) vector.

### Additional information

_No response_

### Operating system

macOS Ventura 13.5

### Matplotlib Version

3.7.1

### Matplotlib Backend

Installed osx event loop hook. MacOSX

### Python version

3.10.11

### Jupyter version

4.0.1

### Installation

conda

Start by running the provided reproduction with matplotlib.pyplot.quiver on a polar Axes. Trace how the polar coordinates and the angles="xy" vectors are interpreted, then verify that the resulting arrows match the requested northward and eastward components. Done means the behavior is corrected or clearly documented, with a regression check for this example.

Written by the indexing model from the issue text.