In June of 1992 a NASA sounding rocket was fired into the Arecibo heater beam to provide in situ observations of artificially induced ionospheric irregularities. In this paper we provide a radar scattering calculation based on in situ data and compare the same with previous remote sensing experiments and with theory. The calculated backscatter cross section is in good agreement with prior observations over the Arecibo heater at 50 MHz. More important, when we scale the observed in situ power spectrum appropriately and compare it with multiradar cross-sectional results from the Platteville, Colorado experiments, we find a remarkably similar radar frequency dependence, albeit one shifted to smaller scales over the higher-latitude site. Even though the rms fluctuation level is almost the same over the Arecibo and Platteville heaters, the shift in scales toward smaller structures over Platteville explains the much larger VHF radar cross section measured there. Comparison of our waveform and its power spectrum with similar predictions from a recent theory shows excellent agreement for k values up to about 5 times the breakpoint in the spectrum of the theoretical prediction. Taken together, these results give very strong evidence for the production of needle-like solitary structures as the dominant final state when high-power radio waves reflect from a magnetized plasma. The organization of these structures by as yet unexplained processes may explain the scales between 10 m and 10 km which occur in the heated volume. Finally, the dominant needle-like field-aligned density depletions seem to support a second source of smaller-scale irregularities. This creates a second break in the power law slope from its one-dimensional value of k-4.3 predicted by theory to one more nearly characterized by k-3. The multiradar results from Platteville show a similar break, and we speculate that a density and/or temperature-gradient-driven instability such as the drift wave is operating. ¿ 1999 American Geophysical Union