@Roy_Smart has helpfully answered the second part of the question, so I'll answer the first :)
TL;DR
Basically NASA just pasted a Rowland circle on the telescope with adjustments and other equipment like microshutters to make FORTIS.
Longer answer:
First, let's have a look at the inside of the telescope:
FORTIS sounding rocket optical layout from Researchgate.net - FORTIS Pathfinder to the Lyman continuum:
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/56KYq.png)
Now this may sound confusing, but there is actually a Rowland circle, just not shown properly in the 2D layout:
Spectroscopic design in the far ultraviolet below Lyα demands a minimalist approach. The “two-bounce” prime focus designs of the HUT, OREFUS and FUSE observatories used relatively fast parabolic primary mirrors feeding the slit of a Rowland circle spectrograph. Although these two-bounce designs satisfy the requirement for efficiency they have poor off-axis spectral imaging performance. This is because fast focal ratios tend to increase the height of the astigmatic image, limiting the spatial resolution perpendicular to the dispersion. Astigmatism control is a major challenge in Rowland circle spectrographs and methods have been developed to eliminate it at select wavelengths, either by controlling the gratingfigure12or with holographic ruling methods.
From here, we can deduce a Rowland circle spectrograph is being used.
Noda first wrote out in detail the general aberration theory for holographic gratings, and since then techniques to minimize spectroscopic aberration shave become increasingly sophisticated. They reached a new level when solutions were found that made it possible to control not only astigmatism but coma at the high ruling densities required for the FUSE mission. The high resolution holographic gratings recorded for the Cosmic Origins Spectrograph (COS – slated to have been flown on the now canceled servicing mission 4 to HST) removed the spherical aberration with an aspherical grating figure along with a FUSE like aberration minimization solution. Recently developed numerically optimized solutions for two-bounce systems achieve excellent narrow band spectroscopic imaging with the grating used off the Rowland circle. An off-axis parabola design has a high enough spectral resolving power to cleanly separate the $O\; VI\; λλ1032 – 1037$ doublet over a 0.5° field of view and yield $4 – 9′′$ of spatial resolution.
Now according to the Researchgate.net - FORTIS Instrument Summary, there are definitely circles:
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/M0AHl.png)
Further proven by the overall design of FORTIS:
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/Mk8yH.png)
Mathematical explanation of Rowland circle
From the previous research paper in a foonote:
In a Rowland circle spectrograph all the rays at wavelengths satisfying the grating equation entering a slit placed on a circle, with diameter is equal to the radius of the concave grating, are diffracted to a tangential focus on the same circle
For reference, here is the grating equation derivation from Newport.com - Diffraction grating physics and Newport.com - The Grating Equation:
From:
![enter image description here](https://cdn.statically.io/img/i.sstatic.net/gS3My.png)
Figure 2-2. Geometry of diffraction, showing planar wavefronts. Two parallel rays, labeled 1 and 2, are incident on the grating one groove spacing d apart and are in phase with each other at wavefront A. Upon diffraction, the principle of constructive interference implies that these rays are in phase at diffracted wavefront B if the difference in their path lengths, $d \sinα + d \sinβ$, is an integral number of wavelengths; this in turn leads to the grating equation. [Huygens wavelets not shown.]
We can deduce:
$$ m\lambda=d_G(\sin\alpha+\sin\beta_m) $$