GX 13+1 shows a larger fractional loss at 2–4 keV than at 4–10 keV near its fixed dip phase in 2017–2026 MAXI data

GX 13+1 has known X-ray dips near a 24.5274-day clock. This work tests that fixed clock in later public MAXI data. It uses MJD 58000–61336, with the end excluded, and the same 33,605 selected 90-minute bins in both energy bands. Each accepted orbit has equal weight. A fixed dip window is compared with two local reference windows. The 114 accepted orbits show a mean fractional loss of 5.67% at 2–4 keV and 2.63% at 4–10 keV. The difference is 3.05 percentage points. A paired bootstrap of eight-orbit calendar groups gives a 95% interval of 0.79–5.25 percentage points. This interval assumes independent, representative groups and excludes unknown calibration errors. Daily data, actual scan times, other phase windows, and clock-error checks give similar energy contrasts. The 2022–2026 subset alone has a contrast interval that includes zero. The result adds a uniform later-record estimate to the known dip behavior. It does not establish a constant dip depth or a change in orbital period. The package includes exact source data, provenance, code, saved bootstrap draws, figures, and reproduction checks.

GX 13+1 · MAXI · X-ray dips · orbital ephemeris · energy dependence · calendar group bootstrap · reproducible astronomy