Measurements of the electric and magnetic form factors of the proton from<mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msup><mml:mrow><mml:mi mathvariant="italic">Q</mml:mi></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math>=1.75 to 8.83 (GeV/<i>c</i><mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" display="inline"><mml:mrow><mml:msup><mml:mrow><mml:mo>)</mml:mo></mml:mrow><mml:mrow><mml:mn>2</mml:mn></mml:mrow></mml:msup></mml:mrow></mml:math>

The proton elastic form factors ${\mathit{G}}_{\mathit{E}\mathit{p}}$(${\mathit{Q}}^{2}$) and ${\mathit{G}}_{\mathit{M}\mathit{p}}$(${\mathit{Q}}^{2}$) have been extracted for ${\mathit{Q}}^{2}$=1.75 to 8.83 (GeV/c${)}^{2}$ via a Rosenbluth separation to ep elastic cross section measurements in the angular range 13\ifmmode^\circ\else\textdegree\fi{}\ensuremath{\le}\ensuremath{\theta}\ensuremath{\le}90\ifmmode^\circ\else\textdegree\fi{}. The ${\mathit{Q}}^{2}$ range covered more than doubles that of the existing data. For ${\mathit{Q}}^{2}$4 (GeV/c${)}^{2}$, where the data overlap with previous measurements, the total uncertainties have been reduced to 14% in ${\mathit{G}}_{\mathit{E}\mathit{p}}$ and 1.5% in ${\mathit{G}}_{\mathit{M}\mathit{p}}$. Results for ${\mathit{G}}_{\mathit{E}\mathit{p}}$(${\mathit{Q}}^{2}$) are consistent with the dipole fit ${\mathit{G}}_{\mathit{D}}$(${\mathit{Q}}^{2}$)=(1+${\mathit{Q}}^{2}$/0.71${)}^{\mathrm{\ensuremath{-}}2}$, while those for ${\mathit{G}}_{\mathit{M}\mathit{p}}$(${\mathit{Q}}^{2}$)/${\mathrm{\ensuremath{\mu}}}_{\mathit{p}}$${\mathit{G}}_{\mathit{D}}$(${\mathit{Q}}^{2}$) decrease smoothly from 1.05 to 0.92. Deviations from form factor scaling are observed up to 20%. The ratio ${\mathit{Q}}^{2}$${\mathit{F}}_{2}$/${\mathit{F}}_{1}$ is observed to approach a constant value for ${\mathit{Q}}^{2}$&gt;3 (GeV/c${)}^{2}$. Comparisons are made to vector meson dominance, dimensional scaling, QCD sum rule, diquark, and constituent quark models, none of which fully characterize all the new data.