
Lowmass white dwarfs (LMWDs) are believed to be exclusive products of binary
evolution, as the Universe is not yet old enough to produce them from single
stars. Because of the strong tidal forces operating during the binary
interaction phase, the remnant host systems observed today are expected to have
negligible eccentricities. Here, we report on the first unambiguous
identification of a LMWD in an eccentric (e=0.13) orbit with a millisecond
pulsar, which directly contradicts this picture. We use our spectra and
radiotiming solution (derived elsewhere) to infer the WD temperature T_eff =
8600 +/ 190 K) and 3D systemic velocity (179.5 km\s). We also place
modelindependent constraints on the WD radius (R_WD = 0.024+/ 0.004/0.002
R_sun) and surface gravity (log g = 7.11 +/ 0.08/0.16 dex). The WD and
kinematic properties are consistent with the expectations for lowmass Xray
binary evolution and disfavour a threebody formation channel. In the case of
the high eccentricity being the result of a spontaneous phase transition, we
infer a mass of 1.6 M_sun for the progenitor of the pulsar, which is too low
for the quarknova mechanism proposed by Jiang et al. (2015). Similarly, the
scenario of Freire & Tauris (2014), in which a WD collapses onto a neutron star
via an rotationallydelayed accretioninduced collapse, requires both a
highmass differentially rotating progenitor and a significant momentum kick at
birth under our constraints. Contrarily, we find that eccentricity pumping via
interaction with a transient circumbinary disk is consistent with all inferred
properties. Finally, we report tentative evidence for pulsations which, if
confirmed, would transform the star into an unprecedented laboratory for WD
physics and stellar convection.

We compute upper limits on the nanohertzfrequency isotropic stochastic
gravitational wave background (GWB) using the 9year data release from the
North American Nanohertz Observatory for Gravitational Waves (NANOGrav)
collaboration. We set upper limits for a GWB from supermassive black hole
binaries under power law, broken power law, and free spectral coefficient GW
spectrum models. We place a 95\% upper limit on the strain amplitude (at a
frequency of yr$^{1}$) in the power law model of $A_{\rm gw} < 1.5\times
10^{15}$. For a broken power law model, we place priors on the strain
amplitude derived from simulations of Sesana (2013) and McWilliams et al.
(2014). We find that the data favor a broken power law to a pure power law with
odds ratios of 22 and 2.2 to one for the McWilliams and Sesana prior models,
respectively. The McWilliams model is essentially ruled out by the data, and
the Sesana model is in tension with the data under the assumption of a pure
power law. Using the broken powerlaw analysis we construct posterior
distributions on environmental factors that drive the binary to the GWdriven
regime including the stellar mass density for stellarscattering, mass
accretion rate for circumbinary disk interaction, and orbital eccentricity for
eccentric binaries, marking the first time that the shape of the GWB spectrum
has been used to make astrophysical inferences. We then place the most
stringent limits so far on the energy density of relic GWs,
$\Omega_\mathrm{gw}(f)\,h^2 < 4.2 \times 10^{10}$, yielding a limit on the
Hubble parameter during inflation of $H_*=1.6\times10^{2}~m_{Pl}$, where
$m_{Pl}$ is the Planck mass. Our limit on the cosmic string GWB,
$\Omega_\mathrm{gw}(f)\, h^2 < 2.2 \times 10^{10}$, translates to a
conservative limit of $G\mu<3.3\times 10^{8}$  a factor of 4 better than the
joint Planck and high$l$ CMB data from other experiments.

We analyze timing noise from five years of Arecibo and Green Bank
observations of the seventeen millisecond pulsars of the NorthAmerican
Nanohertz Observatory for Gravitational Waves (NANOGrav) pulsar timing array.
The weighted autocovariance of the timing residuals was computed for each
pulsar and compared against two possible models for the underlying noise
process. The first model includes red noise and predicts the autocovariance to
be a decaying exponential as a function of time lag. The second model is
Gaussian white noise whose autocovariance would be a delta function. We also
perform a ``nearestneighbor" correlation analysis. We find that the
exponential process does not accurately describe the data. Two pulsars,
J16431224 and J1910+1256, exhibit weak red noise, but the rest are well
described as white noise. The overall lack of evidence for red noise implies
that sensitivity to a (red) gravitational wave background signal is limited by
statistical rather than systematic uncertainty. In all pulsars, the ratio of
nonwhite noise to white noise is low, so that we can increase the cadence or
integration times of our observations and still expect the rootmeansquare of
timing residual averages to decrease by the squareroot of observation time,
which is key to improving the sensitivity of the pulsar timing array.

We derive scaling laws for the signaltonoise ratio of the optimal
crosscorrelation statistic, and show that the large powerlaw increase of the
signaltonoise ratio as a function of the the observation time $T$ that is
usually assumed holds only at early times. After enough time has elapsed,
pulsar timing arrays enter a new regime where the signal to noise only scales
as $\sqrt{T}$. In addition, in this regime the quality of the pulsar timing
data and the cadence become relatively unimportant. This occurs because the
lowest frequencies of the pulsar timing residuals become gravitationalwave
dominated. Pulsar timing arrays enter this regime more quickly than one might
naively suspect. For T=10 yr observations and typical stochastic background
amplitudes, pulsars with residual RMSs of less than about $1\,\mu$s are already
in that regime. The best strategy to increase the detectability of the
background in this regime is to increase the number of pulsars in the array. We
also perform realistic simulations of the NANOGrav pulsar timing array, which
through an aggressive pulsar survey campaign adds new millisecond pulsars
regularly to its array, and show that a detection is possible within a decade,
and could occur as early as 2016.