Magellan: Astronomy & Computational Research

• Magellan is a multifaceted designation that encompasses historical astronomical naming, high-performance telescope infrastructures, and computational frameworks for complex search spaces.
• The Magellan telescopes, notably the twin 6.5 m systems at Las Campanas, deliver superior image quality through advanced adaptive optics, turbulence characterization, and precise instrument calibration.
• Computational frameworks using the Magellan name apply methods like MCMC, MCTS, and evolutionary optimization to efficiently navigate vast and structured parameter or goal spaces.

to=arxiv_search.search արկել อาคารจีเอ็มเอ็มjson {"query":"Magellan arXiv telescope instrument framework site seeing Magellanic Clouds", "max_results": 10, "sort_by":"relevance"} to=arxiv_search.search 大发快三是 】!【json {"query":"(McLeod et al., 2012, Walker et al., 2015, Zhou et al., 2012, McGurk et al., 2020, Layden et al., 22 Jan 2026, Kopon et al., 2010, Cosens et al., 2024, Floyd et al., 2010, Dennefeld, 2020, Arriagada, 2011, Chilingarian et al., 2015, Accomando et al., 2019, Chang, 24 Oct 2025, Gaven et al., 11 Feb 2025, Chen et al., 28 Jan 2026, Kopon et al., 2014)", "max_results": 20, "sort_by":"relevance"} Magellan is a historically and technically layered designation. In astronomy, it is attached both to the southern-sky objects now called the Magellanic Clouds and to the Magellan observing infrastructure at Las Campanas Observatory, where the twin 6.5 metre telescopes support a broad instrumental ecosystem for spectroscopy, adaptive optics, and high-cadence imaging. In contemporary computational research, the same name has also been adopted for several unrelated frameworks in phenomenology, LLM-guided search, autotelic curriculum formation, and compiler optimization (Dennefeld, 2020, McLeod et al., 2012, Accomando et al., 2019, Chang, 24 Oct 2025, Gaven et al., 11 Feb 2025, Chen et al., 28 Jan 2026).

1. Historical and astronomical naming

The strongest historical misconception corrected in the literature is that the Magellanic Clouds were discovered by Ferdinand Magellan or named by Magellan himself. The historical analysis states explicitly that the Clouds were known before Magellan’s voyage, and that neither Magellan nor Antonio Pigafetta coined the modern name. Indigenous South American peoples already had their own interpretations of the objects; the Portuguese used “Clouds of the Cape”; and scientific Latin long preferred Nubecula Minor and Nubecula Major. The modern label “Magellanic Clouds” became standard only at the end of the 19th century, after Latin ceased to dominate scientific communication (Dennefeld, 2020).

This chronology matters because it separates discovery, observation, and later nomenclature. The Strait of Magellan acquired Magellan’s name relatively quickly, within about 20 years of the voyage, because it was a navigational breakthrough of immediate geopolitical value. The Clouds followed a different trajectory: diffuse, less useful for navigation, and embedded in multiple naming traditions before the Magellan association stabilized. The historical record therefore supports an indirect rather than originary connection between Magellan and the Clouds (Dennefeld, 2020).

2. The Magellan telescopes as an observational platform

In modern astronomy, “Magellan” most often denotes the large-telescope infrastructure at Las Campanas Observatory. A central site-characterization result is that the twin 6.5 metre Magellan telescopes generally deliver science image quality better than nearby DIMM seeing would suggest. Over a 1.5 year study using about $10^5$ stars, the median DIMM seeing was $0\farcs625$, while the median Magellan DIPSF FWHM was $0\farcs575$; Magellan outperformed the DIMM 69% of the time, and 98% of the time when DIMM seeing was at least $1\arcsec$ (Floyd et al., 2010).

The same study argues that two effects are decisive. First, finite turbulence outer scale matters for a 6.5 m aperture, with an average effective $\mathcal{L}_0 \sim 25\ \mathrm{m}$. Second, and more strongly, wind speed is the dominant driver of the Magellan–DIMM discrepancy: Magellan outperforms the DIMMs most markedly when the wind is strongest. The work also reports the first detection of a negative DIMM bias, arising when the DIMM is affected by optical aberrations and the turbulence profile is dominated by upper atmospheric layers (Floyd et al., 2010).

A plausible implication is that Magellan’s instrument performance cannot be inferred from site monitors alone. The delivered focal-plane quality is shaped by large-aperture turbulence physics, local wind effects, and instrument-specific optical behavior, which is why several Magellan papers treat calibration, AO correction, and detector performance as system-level rather than component-level problems.

3. Spectroscopic and integral-field instrumentation

The Magellan instrument suite spans cryogenic near-infrared spectroscopy, fiber spectroscopy, and image-slicer integral-field concepts. A representative system is MMIRS, the cryogenic multiple-slit infrared spectrograph/imager designed for the f/5 Cassegrain focus of either the MMT or the Magellan Clay 6.5 m telescope. MMIRS operates over $0.9\text{--}2.4~\mu\mathrm{m}$, covering the near-IR $Y$, $J$, $H$, and $K$ bands, with a HAWAII-2 $0\farcs625$0 detector, a $0\farcs625$1 arcmin imaging field, $0\farcs625$2 arcsec/pixel sampling, long slits up to $0\farcs625$3 arcmin, custom slit masks over $0\farcs625$4 arcmin, and spectral resolution $0\farcs625$5 to $0\farcs625$6. It was commissioned in 2009 at the MMT and has been in routine operation at the Magellan Clay telescope since 2010 (McLeod et al., 2012).

The MMIRS software stack is itself part of the Magellan technical identity. Its stand-alone IDL reduction pipeline supports long-slit and MOS data, uses paired spatial dithers, implements a modified Kelson-style sky model, and reports that sky subtraction on difference spectra reaches the Poisson photon noise limits. Its telluric correction uses a hybrid approach: one observed A0V standard defines an empirical transmission function, which is then differentially transferred with atmospheric transmission models to the science-target airmass. The pipeline was designed for real-time reduction and data-quality control during observations on Magellan Clay (Chilingarian et al., 2015).

Integral-field development has extended Magellan’s capabilities beyond classical slit and fiber modes. ROSIE, the Reformatting Optically-Sensitive IMACS Enhancement IFU for the Magellan Baade Telescope, is an image-slicer IFU for IMACS with a $0\farcs625$7 field of view, pre-sliced into four $0\farcs625$8 subfields and then into 84 slices of $0\farcs625$9. The four slicers produce four pseudo-slits spaced six arcminutes apart across the IMACS f/2 camera field, giving about $0\farcs575$0 and $0\farcs575$1 Å wavelength coverage (McGurk et al., 2020). MIRMOS, by contrast, is a planned next-generation Magellan near-infrared MOS+IFS facility: simultaneous spectroscopy from $0\farcs575$2, imaging from $0\farcs575$3, five channels, a configurable slit unit with 92 pairs of masking bars, and a deployable slicer IFU with $0\farcs575$4 field of view at $0\farcs575$5. Its IFU uses 23 slices and freeform pupil mirrors, and is described as the largest FoV IFS operating at these wavelengths from either the ground or space (Cosens et al., 2024).

System	Primary mode	Key parameters
MMIRS	Cryogenic multi-slit IR spectroscopy/imaging	$0\farcs575$6; $0\farcs575$7 arcmin; $0\farcs575$8
ROSIE	Image-slicer IFU for IMACS	$0\farcs575$9; 84 slices; $1\arcsec$0
MIRMOS	Planned MOS + IFU near-IR spectrograph	$1\arcsec$1; $1\arcsec$2; IFU $1\arcsec$3

Together these systems show a consistent Magellan design pattern: reuse of existing spectrographs where possible, deployment of multiplexing or slicing architectures to increase survey efficiency, and explicit optimization for faint, extended, or redshifted targets.

4. Adaptive optics, visible imaging, and high-speed timing

Visible adaptive optics is a major Magellan theme. The Magellan VisAO camera was developed as the visible-light science instrument for the Magellan Adaptive Secondary AO system on the 6.5 m Magellan Clay telescope. It is designed for $1\arcsec$4 imaging over an $1\arcsec$5 field, using a 585-actuator adaptive secondary mirror, a pyramid wavefront sensor, an advanced atmospheric dispersion corrector, and a high-speed shutter. The stated performance goal is diffraction-limited visible imaging with a resolution of $1\arcsec$6 mas, about $1\arcsec$7 sharper in FWHM than HST at the same wavelength (Kopon et al., 2010).

The enabling technologies are unusually specific. The advanced ADC must correct roughly $1\arcsec$8 of lateral color to better than $1\arcsec$9, and the adopted 2-triplet design is reported as 58% better than a conventional 2-doublet ADC at large zenith angle. The system also addresses the strong temporal variability of visible Strehl: at $\mathcal{L}_0 \sim 25\ \mathrm{m}$0, simulations show fluctuations on $\mathcal{L}_0 \sim 25\ \mathrm{m}$1 ms timescales, with Strehl ranging from 50% down to 10% over several hundred milliseconds. Magellan’s answer is not only fast CCD operation but a fast asynchronous shutter, combined with telemetry-based prediction of good and bad Strehl intervals (Kopon et al., 2010).

Closed-loop validation further defined the system. The 2014 status paper reports 85% Strehl at $\mathcal{L}_0 \sim 25\ \mathrm{m}$2 nm in the fully integrated Arcetri tower test, 37% Strehl in $\mathcal{L}_0 \sim 25\ \mathrm{m}$3-band with 400 controlled modes under simulated turbulence with $\mathcal{L}_0 \sim 25\ \mathrm{m}$4 cm at 550 nm, and stable ASM operation in closed loop at 1 kHz. The same paper details simultaneous visible/IR AO operation via a dichroic beamsplitter, with visible light sent to the W-unit and infrared light directed to CLIO or BLINC/MIRAC4 (Kopon et al., 2014).

High-speed optical imaging extends this AO/time-domain axis. proto-Lightspeed, commissioned on the Nasmyth East port of the Magellan Clay Telescope, uses the Hamamatsu ORCA-Quest 2 qCMOS camera to deliver deep sub-electron read noise, a $\mathcal{L}_0 \sim 25\ \mathrm{m}$5 diameter field at up to 200 Hz, and windowed imaging up to 6600 Hz for a $\mathcal{L}_0 \sim 25\ \mathrm{m}$6 field. It provides adjustable pixel scales from $\mathcal{L}_0 \sim 25\ \mathrm{m}$7 to $\mathcal{L}_0 \sim 25\ \mathrm{m}$8, seeing-limited performance in $\mathcal{L}_0 \sim 25\ \mathrm{m}$9, $0.9\text{--}2.4~\mu\mathrm{m}$0, and $0.9\text{--}2.4~\mu\mathrm{m}$1, and absolute on-sky timing accuracy better than $0.9\text{--}2.4~\mu\mathrm{m}$2. It is explicitly positioned as a prototype for a future five-channel facility instrument, Lightspeed, intended for simultaneous $0.9\text{--}2.4~\mu\mathrm{m}$3 imaging over a $0.9\text{--}2.4~\mu\mathrm{m}$4 field (Layden et al., 22 Jan 2026).

The Magellan name also extends to giant-segment phasing in the GMT context. A 2025 demonstration of the holographic dispersed fringe sensor addresses differential piston in segmented pupils, a core problem for the Giant Magellan Telescope. The HDFS provides approximately $0.9\text{--}2.4~\mu\mathrm{m}$5 dynamic piston range and, in the reported first on-sky use, phased each segment to within $0.9\text{--}2.4~\mu\mathrm{m}$6 residual piston WFE at $0.9\text{--}2.4~\mu\mathrm{m}$7 nm, corresponding to about $0.9\text{--}2.4~\mu\mathrm{m}$8 nm RMS residual piston WFE across the aperture in poor seeing (Kautz et al., 14 Jan 2025).

5. Representative scientific programs on Magellan

The observational literature shows that Magellan is not tied to a single astrophysical domain. In exoplanet transmission spectroscopy, MIKE observations of WASP-17b on the 6.5 m Magellan II (Clay) Telescope detected sodium absorption in the Na I D region with an excess $0.9\text{--}2.4~\mu\mathrm{m}$9 per cent transit signal at 1.5 Å bandwidth and 4.5$Y$0 confidence. A distinctive methodological feature was the use of nearby interstellar Na I absorption as an internal reference for residual systematics. The result independently confirmed earlier VLT evidence for a highly inflated sodium-bearing atmosphere (Zhou et al., 2012).

In resolved stellar-population work, Magellan/M2FS spectroscopy of Reticulum 2 used HiRes mode over $Y$1 Å at $Y$2, with 1.2 arcsec fibers over a 29 arcmin field. Of 182 observed stellar targets, 38 had spectra sufficient for simultaneous estimation of $Y$3, $Y$4, $Y$5, and $Y$6, and 18 were confirmed as members. The paper interprets Reticulum 2 as a bona fide galaxy rather than a star cluster, while explicitly noting that the dynamical inference depends on dynamical equilibrium and negligible contamination from binary stars (Walker et al., 2015).

Magellan planet-search infrastructure has also been used to construct stellar-noise priors for Doppler work. A chromospheric activity survey from the Magellan Planet Search Program derived Ca II H & K S-values and $Y$7 measurements for $Y$8 southern F, G, K, and M main-sequence stars from $Y$9 archival MIKE spectra, then used those indices to estimate radial-velocity jitter, rotation periods, and ages. The practical purpose was target ranking for the New Magellan Planet Search aimed at rocky and habitable planets (Arriagada, 2011).

Near-infrared work with MMIRS was designed for a comparably broad range of science, from exoplanet atmospheres to Ly$J$0 emitters. Its multiplexed slit-mask capability is explicitly described as well matched to the surface density of high-redshift galaxies, particularly where rest-frame optical lines such as H$J$1, [O III], [O II], and H$J$2 are shifted into the near-IR for galaxies at $J$3 (McLeod et al., 2012).

6. “Magellan” as a family of computational frameworks

Outside telescope instrumentation, “Magellan” has become a recurrent name for search and inference systems. In phenomenology, the framework introduced in “Voyage Across the 2HDM Type-II with Magellan” is an MCMC-driven global-analysis environment built around T3PS, HiggsSignals, HiggsBounds, 2HDMC, and SusHi. Its stated methodological distinction is that it retains full point-by-point information from a high-dimensional scan, so projected 2D views preserve hidden correlations rather than relying only on scan-and-project maximization (Accomando et al., 2019).

In LLM-based creativity research, “Magellan” denotes a guided MCTS framework for scientific idea generation. It constructs a semantic compass from embeddings of 16,582 paper abstracts, then combines long-range vector guidance with a landscape-aware value function balancing coherence, novelty, and progress. In the reported evaluation, it outperformed Zero-shot, CoT, ReAct, ToT, AI Scientist, and SciPip, achieving Plausibility 8.98, Clarity 9.30, Innovation 8.54, Overall 8.94, and a 92.0% win rate (Chang, 24 Oct 2025).

A second LLM-agent use, styled in all caps as MAGELLAN, concerns metacognitive prediction of learning progress in very large goal spaces. The system estimates current and past competence from the LLM’s latent goal representation, computes absolute learning progress, and samples goals proportionally to that estimate. In the Little-Zoo benchmark, whose full goal space is about 19.5 million goals, the paper reports that MAGELLAN was the only method allowing the agent to fully master a large and evolving goal space (Gaven et al., 11 Feb 2025).

A third computational usage appears in compiler optimization. There, Magellan is an AlphaEvolve-based framework that synthesizes executable C++ compiler heuristics, compiles them into existing systems, evaluates them on macro-benchmarks, and refines them with evolutionary search plus autotuning. Reported outcomes include 5.23% size improvement over LLVM’s human-developed inlining heuristic after a 1.5-day search, more than 5% size reduction with autotuning in about 5 hours, a register-allocation policy matching intricate human-designed heuristics, and a 7% improvement on an XLA graph-rewriting task (Chen et al., 28 Jan 2026).

These computational usages are homonymous rather than genealogically connected. What they share is methodological emphasis on guided exploration of large search spaces—BSM parameter manifolds, latent conceptual spaces, autotelic goal spaces, or compiler-policy spaces—rather than any direct link to the Magellan telescopes.

7. Conceptual unity and scope

Across these literatures, the name “Magellan” functions less as a single object than as a stable label for exploration under difficult observational or combinatorial constraints. In historical astronomy, it marks a retrospective naming process rather than a direct discovery. In observatory practice, it denotes a high-performance platform whose strengths include near-IR multiplexing, visible AO, precision timing, and continuing IFU development. In algorithmic research, it names frameworks that use MCMC, MCTS, competence prediction, or evolutionary synthesis to navigate complex spaces efficiently (Dennefeld, 2020, Floyd et al., 2010, McLeod et al., 2012, McGurk et al., 2020, Cosens et al., 2024, Layden et al., 22 Jan 2026, Accomando et al., 2019, Chang, 24 Oct 2025, Gaven et al., 11 Feb 2025, Chen et al., 28 Jan 2026).

This suggests a useful editorial shorthand: “Magellan” in current research is a multi-domain signifier of exploration. In astronomy that exploration is literal—surveying the southern sky, resolving diffraction-limited structure, or phasing segmented apertures. In computation it is abstract—searching model space, latent semantic space, open-ended curricula, or optimization heuristics. The commonality is not institutional continuity but recurrent association with technical navigation through spaces that are large, structured, and only partially accessible.