The same as the left but for Cu+Cu collisions. The systematic errors are shown as thin lines with wide caps at the ends and statistical errors are shown as thick lines with small caps at the end. Statistical and systematic errors are very small.

The difference of charge-independent (CI) v2{2} and like-sign (LS) $v_2\{2\}$ for Au+Au and Cu+Cu collisions at 200 (top panel) and 62.4 (bottom panel) GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.The statistical errors are smaller than the marker size and not visible for most of the data.

The difference of charge-independent (CI) v2{2} and like-sign (LS) $v_2\{2\}$ for Au+Au and Cu+Cu collisions at 200 (top panel) and 62.4 (bottom panel) GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.The statistical errors are smaller than the marker size and not visible for most of the data.

The difference of charge-independent (CI) v2{2} and like-sign (LS) $v_2\{2\}$ for Au+Au and Cu+Cu collisions at 200 (top panel) and 62.4 (bottom panel) GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.The statistical errors are smaller than the marker size and not visible for most of the data.

The difference of charge-independent (CI) v2{2} and like-sign (LS) $v_2\{2\}$ for Au+Au and Cu+Cu collisions at 200 (top panel) and 62.4 (bottom panel) GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.The statistical errors are smaller than the marker size and not visible for most of the data.

The LS and CI four-particle cumulant $v_2\{4\}^4$ for Au+Au collisions at 200 and 62.4 GeV for $0.15 < pT < 2.0$ GeV/$c$. The systematic errors are shown as narrow lines with wide caps at the end and statistical errors are shown as thick lines with narrow caps at the end. Statistical errors are not visible for most of the points.

The LS and CI four-particle cumulant $v_2\{4\}^4$ for Au+Au collisions at 200 and 62.4 GeV for $0.15 < pT < 2.0$ GeV/$c$. The systematic errors are shown as narrow lines with wide caps at the end and statistical errors are shown as thick lines with narrow caps at the end. Statistical errors are not visible for most of the points.

The LS and CI four-particle cumulant $v_2\{4\}^4$ for Cu+Cu collisions at 200 and 62.4 GeV for $0.15 < p_T < 2.0$ GeV/c. The most central points (two points for Cu+Cu 62.4 GeV) gives $v_2\{4\}^4 < 0$ for all the data sets. The negative values are probably due to large fluctuations in agreement with Eq. (1). These may include contributions from impact parameter spread and finite multiplicity bin width.

The LS and CI four-particle cumulant $v_2\{4\}^4$ for Cu+Cu collisions at 200 and 62.4 GeV for $0.15 < p_T < 2.0$ GeV/c. The most central points (two points for Cu+Cu 62.4 GeV) gives $v_2\{4\}^4 < 0$ for all the data sets. The negative values are probably due to large fluctuations in agreement with Eq. (1). These may include contributions from impact parameter spread and finite multiplicity bin width.

The difference of charge-independent (CI) $v_2\{4\}$ and like-sign (LS) $v_2\{4\}$ for Au+Au collisions at 200 and 62.4 GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.

The difference of charge-independent (CI) $v_2\{4\}$ and like-sign (LS) $v_2\{4\}$ for Au+Au collisions at 200 and 62.4 GeV vs. the log of $\langle dN_{ch}/d\eta\rangle$.

(Left) The difference between $v_2\{2\}^2$ and $v_2\{4\}^2$ for 200 GeV Au+Au and Cu+Cu collisions for both LS and CI combinations.

(Left) The difference between $v_2\{2\}^2$ and $v_2\{4\}^2$ for 200 GeV Au+Au and Cu+Cu collisions for both LS and CI combinations.

(Right) The difference between $v_2\{2\}^2$ and $v_2\{4\}^2$ for 62.4 GeV Au+Au and Cu+Cu collisions for both LS and CI combinations. The statistical and systematic errors are shown as in previous figures.

(Right) The difference between $v_2\{2\}^2$ and $v_2\{4\}^2$ for 62.4 GeV Au+Au and Cu+Cu collisions for both LS and CI combinations. The statistical and systematic errors are shown as in previous figures.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Au+Au collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models. The upper limit is found using the LS results for $v_2\{2\}$. Data are from the range $0.15 < p_T < 2.0$ GeV/$c$. The shaded bands reflect the uncertainties on the models which are dominated by uncertainty on the distribution of nucleons inside the nucleus. The uncertainty is only shown for the MCG-N and fKLN-CGC models. The uncertainty on the MCG-Q model is the same as for the MCG-N model but is not shown for the visual clarity.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Au+Au collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models. The upper limit is found using the LS results for $v_2\{2\}$. Data are from the range $0.15 < p_T < 2.0$ GeV/$c$. The shaded bands reflect the uncertainties on the models which are dominated by uncertainty on the distribution of nucleons inside the nucleus. The uncertainty is only shown for the MCG-N and fKLN-CGC models. The uncertainty on the MCG-Q model is the same as for the MCG-N model but is not shown for the visual clarity.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Au+Au collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models. The upper limit is found using the LS results for $v_2\{2\}$. Data are from the range $0.15 < p_T < 2.0$ GeV/$c$. The shaded bands reflect the uncertainties on the models which are dominated by uncertainty on the distribution of nucleons inside the nucleus. The uncertainty is only shown for the MCG-N and fKLN-CGC models. The uncertainty on the MCG-Q model is the same as for the MCG-N model but is not shown for the visual clarity.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Au+Au collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models. The upper limit is found using the LS results for $v_2\{2\}$. Data are from the range $0.15 < p_T < 2.0$ GeV/$c$. The shaded bands reflect the uncertainties on the models which are dominated by uncertainty on the distribution of nucleons inside the nucleus. The uncertainty is only shown for the MCG-N and fKLN-CGC models. The uncertainty on the MCG-Q model is the same as for the MCG-N model but is not shown for the visual clarity.

The STAR data compared to PHOBOS data [34] on $\sigma_{v_2}/\langle v_2 \rangle$ with $\delta_2$ for $\Delta\eta > 2$ taken to be zero (see Fig. 6 from Ref. [34]). The shaded band shows the errors quoted from Ref. [34].

The STAR data compared to PHOBOS data [34] on $\sigma_{v_2}/\langle v_2 \rangle$ with $\delta_2$ for $\Delta\eta > 2$ taken to be zero (see Fig. 6 from Ref. [34]). The shaded band shows the errors quoted from Ref. [34].

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Cu+Cu collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Cu+Cu collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Cu+Cu collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models.

The upper limit on $\sigma_{v_2}/\langle v_2 \rangle$ for 200 GeV (left) and 62.4 GeV (right) Cu+Cu collisions from Eq. (9) compared to $\sigma_\varepsilon/\varepsilon$ from Eq. (10) for three different models.

$\rho^0$ production cross section determined with the data set collected with trigger B, as a function of momentum transfer $t$, fitted with a double exponential fit function in the range $t < 0.1$. The statistical errors are shown. The ratio of incoherent to coherent cross sections is measured to be $0.20 \pm 0.8$. The fit parameters are shown in Table I.

Comparison of theoretical predictions to the measured total cross section for coherent $\rho^0$ production as a functionof $\sqrt{s_{NN}}$ . The measured cross section is based on the dataset collected with trigger B. The error bars show the sum of the statistical and systematic uncertainties. See text for details.