Why Does One Wall Read as Two Colours in Daylight and Lamplight?

The paint is fine. A wall reads as two colours when two different lights land on it at once - daylight through the glass at one end, the wall light at the other - and paint can only hand back the light it is given.

The gap between those two lights is wider than it feels. CIE standard illuminant A, the stand-in for incandescent light, is defined at 2856 K; D65, average noon sky daylight, at 6500 K; and north-sky daylight was standardised separately as D75 at 7500 K (X-Rite's illuminant reference). The Houseland bedside sconce carries its light temperature as a variant on the listing, Warm White 2700-3500K or Cool White 5500-7000K, so the warm end of that spread is a decision you make at checkout rather than a property welded into the fitting.

Why do daylight and a wall light show two colours on one wall?

Two sources are sharing one surface and neither of them is winning. The US Department of Energy's fact sheet on LED colour characteristics states it without hedging: objects are not inherently coloured but reflect different proportions of radiant energy, so "if the light incident on a surface changes, the apparent color of the object may change". The wall genuinely is two colours right now. It just is not two paints.

Do the arithmetic on your own room. A 2,700 K bulb burning against a north window standardised at 7500 K is a spread of about 4,800 K across one flat plane, with a soft gradient through the middle where the two overlap in whatever ratio the geometry hands you. That gradient is the seam you keep staring at, and it has no edge because neither source has an edge.

Real daylight ranges further than the standards imply. A two-year run of 2,600 daylight spectra recorded for all sky states at Granada, Spain spanned correlated colour temperatures from 3,758 K to 34,572 K, which is why the seam looks one way on a clear morning and another on a flat grey afternoon. The lamp holds still; the sky does not.

Why don't your eyes correct it the way they do after dark?

Your visual system renormalises to one white point, and you are asking it for two. That renormalising is chromatic adaptation, and it is fast: Fairchild and Reniff tracked it at constant luminance and found adaptation 90% complete after approximately 60 s (JOSA A, 1995). It is the reason a room lit only at 2,700 K stops looking yellow a minute after you walk in.

Even at full strength that compensation stops well short of 100%. Fairchild's 2020 vK20 paper fits degree-of-adaptation values that range from 0.6 to 0.8 with a mean of 0.71 - about 70% complete adaptation, not the total correction people assume they have. The single-source case already leaves a residue you never notice. The two-source case leaves one you cannot stop noticing.

The two-source case has been measured on its own. Yang and Shevell lit a scene from two opposite corners with an opaque dividing wall between the sources, and colour constancy fell whenever a region carried cues to both illuminants (Journal of Vision, 2003). Cues from the second light are not merely useless to your eye - they are actively misleading, and your bedroom wall is running that experiment every winter afternoon.

How do you prove it is the light and not two batches of paint?

Remove the variables one at a time, because a lighting seam moves and a roller seam does not.

1. Switch the wall light off in daylight and wait a full minute before judging. 60 s is roughly where adaptation reaches 90%, so a verdict formed in the first five seconds is your own eye still travelling.
2. Look again after dark with the lamp as the only source. Two ends that match each other at 10 pm and disagree at 3 pm are one paint under two illuminants.
3. Slide a sheet of white printer paper flat along the wall, window end to lamp end. Paper you know is a single colour will shift at exactly the point the wall shifts.
4. Photograph both ends inside one frame with white balance locked. Two separate shots let the phone renormalise each one and quietly delete the evidence.
5. Read the shape of the seam. A batch line is straight, vertical and still there at midnight; a lighting seam is soft, sits diagonally across the wall and travels with the sun.

Painters hit the same trap from the other direction: a sample board approved under shop lighting and hated at home. Our note on why wall art arrives looking different from the photo is this mechanism working on a printed surface.

Which fixes actually work, and which ones only sound like they do?

Only one of them ends the argument: stop lighting that wall with two sources.

Aiming beats swapping bulbs and it is free on any fixture with a moving head - the same trick behind getting the beam onto a book instead of a partner's face, which we worked through in the piece on height and angle for reading in bed. Point the shade along the wall towards the room rather than flat into the plaster and the graze that exaggerates the gradient largely disappears.

Relocating a hardwired fitting is electrical work, not decorating. In England and Wales moving or adding a wall box falls under Part P of the Building Regulations; in the United States the same job follows the NEC as your local authority has adopted it. Have a qualified electrician do it in either market. If the sconce runs on a plug and flex, the move is yours to make and only the trailing cable needs solving - the cord-shortening options are here.

Does matching the bulb's Kelvin to the window close the gap?

Partly, and only for the hours you matched. The Houseland sconce lists two temperatures - Warm White 2700-3500K or Cool White 5500-7000K - on an E27 screw socket, so the source is a choice you can revisit for a few dollars rather than a sealed part of the fixture. Cool White at the top of that band lands close to D65 at 6500 K, which is the honest option for a room used in daylight.

Matching the number on the box is not the same as matching the light. The DOE fact sheet is blunt about the limits of the metric: CCT "characterizes the appearance of the emitted light, not the color of illuminated objects", and "two sources with the same CRI value can render colors very differently". A 6500 K LED and a 6500 K sky can still push the same paint in different directions, because the spectra behind those identical numbers are not the same shape.

There is a second seam nobody mentions, and it survives every bulb swap. The wall is brighter near the lamp than it is at the far end of the same wall, so even with the colour matched you are left with a luminance gradient. That one reads as depth rather than as a wrong colour, which is why most people stop seeing it once the colour difference is gone.

From the shop

Bedside Wall Sconce, Adjustable Wood E27 Wall Light, 6 Colors

E27 socket on a wood round backplate, 8.7 in (22 cm) high with 8.3 in projection. Shade colours are Gray, Green, Blue, Pink, White or Yellow, and the light temperature is listed as Warm White 2700-3500K or Cool White 5500-7000K.

View the adjustable bedside sconce — 6 shade colours & current price

What if the window is north-facing and the afternoon is grey?

Then accept a permanent gap and choose your hours instead. North sky is the far end of the domestic range - the CIE standardised it as D75 at 7500 K - and measured daylight runs far past that. A warm bulb at 2,700 K is around 4,800 K away from it, and no amount of standing there will reconcile the two.

Do not reach for the old rule that warm light must be dim to be comfortable. Fotios's 2017 review in LEUKOS put the Kruithof curve against nine methodologically credible replication studies: five reject it, one gives partial support, three are unclear and none support it. What does survive is an illuminance effect - interiors below roughly 300 lux get rated unpleasant and around 500 lux is enough - while Boyce and Cuttle found that "once the subject is fully adapted to the conditions, the CCT of good colour rendering lamps in the range 2700K to 6300K has little effect on people's impressions of the lighting of the room".

Colour temperature is a preference once you have adapted, so pick the Kelvin you like for the hours the room is actually used, and judge the wall then. Boyce and Cuttle allowed observers 15 to 20 minutes to settle before they asked anything - a whole room takes considerably longer than the 60 s that a single patch of wall needs.

What I'd do

Move the light before touching the paint. The seam exists because one plane is carrying two white points, and every fix that leaves both sources aimed at that plane is a compromise dressed up as a solution.

If moving it is not on the table, aim it. The Houseland fixture stands 8.7 in (22 cm) high with 8.3 in projection, so it sits close in and grazes the plaster behind it - swinging the shade out towards the room instead of flat against the wall takes the strongest part of the wash off the surface you are comparing to the window. Anything in our wall lamps collection with a moving head can do the same thing.

Then set expectations honestly. Choose the bulb for the hours you use the room - warm for a bedroom you mostly see after dark, cool for a room you work in at 2 pm - stand there for a full minute before you decide anything, and stop auditing the wall at the seam. On a north-facing winter afternoon, a 2,700 K lamp and a 7500 K sky will never agree, and the paint has done nothing wrong.