The Missing Temporal Crescent
KLARA LANDAU, MD., WERNER WICHMANN, MD., AND ANTON VALAVANIS, M.D.

• PURPOSE: We studied clinically the representa­
tion of the monocular temporal crescent in the
human visual cortex and noted the importance of
using the perimetric techniques best suited to
detect this visual field defect and to study patients
in whom the temporal crescent is missing.
• METHODS: Goldmann perimetry and high-reso­
lution magnetic resonance imaging were per­
formed in two patients with vascular lesions locat­
ed in the anterior striate cortex.
• RESULTS: A monocular visual field defect, the
missing temporal crescent, was found on the side
contralateral to the lesion.
• CONCLUSIONS: The perimetric-magnetic reso­
nance imaging correlation is in exquisite agree­
ment with recent information about the represen­
tation of the visual field in the human primary
visual cortex. Reports of this specific perimetric
finding are rare, in part because of underdetection
with currently used perimetric techniques that
concentrate on the central 30 degrees of the visual
field.

R

ETROCHIASMAL LESIONS OF THE VISUAL PATHways affect the visual fields in a homonymous
fashion. The only exception to this rule is the
missing temporal crescent of one eye, for which no
corresponding homonymous visual field defect can
occur in the fellow eye. The nasal visual field extends
to roughly 60 degrees on the horizontal meridian, as
opposed to a 100-degree extension of the temporal

Accepted for publication Aug. 29, 1994.
From the Departments of Ophthalmology (Dr. Landau) and Neuroradiology (Drs. Wichmann and Valavanis), University Hospital, Zurich,
Switzerland. Data from this study were presented as a poster at the
20th annual North American Neuro-Ophthalmology Society meeting
in Durango, Colorado, February 27-March 3, 1994.
Reprint requests to Klara Landau, M.D., Department of Ophthal­
mology, University Hospital Zurich, Frauenklinikstrasse 24, 8091 Zur­
ich, Switzerland; fax: 411-255-4438.

VOL.119, No. 3

visual field. This well-known phenomenon of a
monocular visual field defect caused by a retrochiasmal lesion is described theoretically in countless
textbooks, but actual case studies are exceedingly rare.
The last series was reported by Bender and Strauss1 in
1937, and a single case was published by Walsh2 in
1974- We studied two patients with purely monocular
visual field defects involving the temporal crescent of
one eye. Magnetic resonance imaging showed vascu­
lar lesions in the contralateral visual cortex.

CASE REPORTS
• CASE l: A 47-year-old woman had a ten-year
history of migraine with visual auras occurring re­
peatedly in her right visual field. Over the years the
frequency of the attacks increased. Computed tomog­
raphy disclosed an arteriovenous malformation in the
left parieto-occipital lobe, and the patient was referred
for treatment. In May 1992 an embolization was per­
formed, and the arteriovenous malformation was
partially occluded. In November 1992 the patient was
referred for neuro-ophthalmic examination. She gave
a history of zigzag scintillations occurring in her right
visual field, which were followed by headaches. Be­
tween such attacks there were no visual symptoms.
The patient's uncorrected visual acuity was 20/20
in each eye. Her pupils were equal and showed no
afferent or efferent defect. Results of slit-lamp exami­
nation were normal. Goldmann perimetry of the right
eye showed a peripheral defect, more pronounced in
the upper than in the lower temporal quadrant,
extending between 60 and 90 degrees along the
horizontal meridian (Fig. 1). The visual field of the
left eye was full, extending to 90 degrees temporally
and to 60 degrees nasally. A dilated fundus examina­
tion was performed in both eyes, with special atten­
tion to the nasal retinal periphery of the right eye. No

© AMERICAN JOURNAL OF OPHTHALMOLOGY 119:345-349

345

120

105

90

75

60

120

90

105

75

άθ

270

RIGHT

Fig. 1 (Landau, Wichmann, and Valavanis). Case 1. Right temporal peripheral defect, more pronounced in the upper
than in the lower quadrant. No homonymous defect on the left side.

abnormality was found. Magnetic resonance imaging
showed a large left temporo-occipital arteriovenous
malformation (Fig. 2).
• CASE 2: A 27-year-old woman was seen at the
Department of Ophthalmology for an emergency
examination on May 28, 1993. Her medical history
was remarkable for migraine headaches with visual
aura since the age of 15 years. She was not taking any
medication except for oral contraceptives. Two days
before the examination the patient had an unusually
severe attack of headache accompanied by nausea
and photophobia. She had to lie down with closed
eyes for several hours. Upon arising she immediately
noted that the most peripheral part of the visual field
on her right was missing. When this field defect
persisted for the next two days, the patient consulted
her general practitioner and was referred for a neuroophthalmologic examination.
Visual acuity was 20/15 in each eye, with slight
astigmatic correction. Pupils were equal and showed
no afferent or efferent defect. Slit-lamp examination
was normal. Goldmann perimetry on the right
showed a peripheral defect, more pronounced in the
lower than in the upper temporal quadrant but
346

otherwise almost identical to the one seen in Case 1
(Fig. 3). The visual field of the left eye was full.
Results of a dilated fundus examination were normal
in both eyes. A 30-degree full-threshold static perime­
try performed ten days after the initial examination
was full. Computed tomography was normal, but on
magnetic resonance imaging performed one month
after onset of symptoms, a small enhancing lesion in
the left upper anterior occipital cortex was seen (Fig.
4). It was interpreted as a subacute stroke. An
incidental pineal cyst was also present.

DISCUSSION
IN 1937 BENDER AND STRAUSS1 PUBLISHED A SERIES OF

ten cases of missing temporal crescent found among
100 perimetrically studied patients with brain tumors
involving the optic radiations. On the basis of these
ten cases, they concluded that temporal crescentic
defects result from involvement of the distal optic
radiation or anterior visual cortex and that they may
represent the earliest sign of brain tumors. They
postulated that "indeed, this defect may even be
found to be the most frequent type of distortion in

AMERICAN JOURNAL OF OPHTHALMOLOGY

M A R C H 1995

Fig. 2 (Landau, Wichmann, and Valavanis). Case 1. Left, Coronal 4-mm T^weighted magnetic resonance image at the
level of the anterior visual cortex, showing a left-sided arteriovenous malformation, draining by dilated veins. For
orientation, an arrowhead marks the contralateral calcarine fissure containing a small vein. Right, T^weighted
parasagittal view of the arteriovenous malformation, located predominantly in the anterior part of the occipital lobe.

the fields of vision of patients with tumor of the
brain." This prediction was not confirmed in subsequent decades. Reports about the missing temporal
crescent are rare. Authors of perimetry books and

120

105

90

75

60

chapters describe the interesting phenomenon of the
missing temporal crescent either entirely on theoretical grounds or in reference to Bender and Strauss.1 In
1974 Walsh2 studied two patients, one with a pre-

120

LEFT

105

W

B

60

RIGHT

Fig. 3 (Landau, Wichmann, and Valavanis). Case 2. Right temporal peripheral defect, more pronounced in the lower
than in the upper quadrant. No homonymous defect on the left side.
VOL. i i 9,

No. 3

MISSING TEMPORAL CRESCENT

347

Fig. 4 (Landau, Wichmann, and Valavanis). Case 2. Three-millimeter Tj-weighted gadolinium-enhanced magnetic
resonance image, showing a minute subacute cortical infarction. Left, The axial view demonstrates that the lesion is
located in the most anterior portion of the occipital cortex, measuring approximately 13 X 8 mm. The distance between
the bars on the right is 10 mm. Right, The coronal view demonstrates that only the cortex above the calcarine fissure is
involved.

served monocular temporal crescent in the presence
of a dense homonymous hemianopsia and the second
with loss of the temporal crescent. Neuroradiologic
studies were not available at that time. Clinicopathologic correlation was provided in a single case by
Ask-Upmark 3 in 1932.
Our first patient demonstrates that large long­
standing arteriovenous malformations may produce
few neurologic deficits. O n the basis of the neuroradi­
ologic findings, this patient might have had any visual
field defect, ranging from total right homonymous
hemianopsia to no deficit at all. The missing temporal
crescent in this patient may be a result of damage to
optic radiation, cortical damage, or a combination of
both.
Our second patient is of particular interest because
the missing temporal crescent was caused by a purely
cortical lesion. In a recent study by Horton and
Hoyt,4 in which a revised map of the representation of
the visual field in human striate cortex was provided,
the 40 degrees of the temporal crescent correspond to
less than 10% of the total surface area of the striate
cortex. This surface area is roughly equal to the area
corresponding to the central 1 degree of vision.
Horton and Hoyt4 stated that testing the central 30
348

degrees of visual field by using full-threshold automat­
ed perimetry provides information on as much as 83%
of the striate cortex. In the striate cortex, the most
anterior portion that subserves the temporal crescent
and gets input only from the contralateral eye has a
distinct feature: it lacks ocular dominance columns.
This feature allowed precise identification and mea­
surement of this cortical region. The cortex subserv­
ing the monocular crescent in the case by Horton
and associates5 was found to have a dimension of 154
mm2. The purely cortical lesion in Case 2 measured
roughly 13 X 8 mm, thus affecting an area of
approximately 100 mm2. If one takes into account
that the temporal crescent was completely missing in
the lower quadrant but was partially spared in the
upper quadrant, then this discrepancy in size is easily
reconciled. The magnetic resonance scan shows that
the lesion involved the upper calcarine bank, which
represents the lower temporal crescent. Thus, both
the size and the location of the cortical lesion
correspond to the perimetric defect and fit nicely with
the revised map 4 of the representation of the visual
field in the human striate cortex, as discussed previ­
ously.
Various reasons may account for the paucity of

AMERICAN JOURNAL OF OPHTHALMOLOGY

MARCH 1995

reports on the missing temporal crescent. Such a
defect may simply be unnoticed by the patient. In
congenital lesions, even total homonymous hemianopsias can remain undetected for many years.6 It is
thus not surprising that our first patient was asympto­
matic. In acquired lesions, the missing temporal
crescent may be more readily appreciated if it occurs
suddenly, as in Case 2 and in the patient studied by
Walsh.2 In slowly progressive lesions it may remain
unnoticed by the patient because of little resulting
handicap.
This type of visualfielddefect can be missed by the
examiner. When performing Goldmann perimetry,
the largest stimulus should be presented at the edge of
the cupola, to include the temporal crescent. Another
quick way to examine the temporal crescent is to
move the examiner's hand slowly, starting behind the
patient's head, anteriorly. This approach will show
that the temporal visual field border extends beyond
90 degrees, with variations because of individual
anatomic factors, such as deep or shallow orbits.
Confrontation testing or finger counting in the four
quadrants will not include all of the temporal periph­
eral field because the examiner's hand is not long
enough to enter the temporal crescent out of the
blind area. Computerized automated perimetry will
not show a missing temporal crescent.
On the basis of a prospective study of 215 visual
field defects caused by lesions of the visual pathways,
Wellings7 stated that such lesions are detectable
within 30 degrees of fixation and that moreperipheral field testing is ineffective. We agree with
Horton and Hoyt4 that a lesion affecting only the
most anterior 10% of the striate cortex must be truly
rare. For most clinical situations it is sufficient to
examine the central visual field when a retrochiasmal

VOL.119, No. 3

lesion is suspected. However, the perimetrist should
be aware that, although a 30-degree visual field
provides information on 83% of the striate cortex, it
does not provide information about the remaining
17%.
With this study, we draw attention to the missing
temporal crescent, which may be a more common
defect than is documented in the medical literature.
A simple finger confrontation screening for such a
defect will quickly enable the examiner to decide
whether careful Goldmann perimetry should follow.
We provide a magnetic resonance image of a missing
temporal crescent caused by an isolated cortical
lesion. It is in accord with the current information on
the representation of the visual field in the human
striate cortex.4,8

REFERENCES
1. Bender MB, Strauss I. Defects in visualfieldof one eye only in
patients with a lesion of one optic radiation. Arch Ophthalmol
1937;17:765-87.
2. Walsh TJ. Temporal crescent or half-moon syndrome. Ann
Ophthalmol 1974;6:501-5.
3. Ask-Upmark E. On the cortical projection of the temporal
half-moon of the visual field. Acta Ophthalmol 1932;10:27190.
4. Horton JC, Hoyt WF. The representation of the visualfieldin
human striate cortex. Arch Ophthalmol 1991;109:816-24.
5. Horton JC, Dagi LR, McCrane EP, de Monasterio FM.
Arrangement of ocular dominance columns in human visual
cortex. Arch Ophthalmol 1990;108:1025-31.
6. Bajandas FJ, McBeath JB, Smith JL. Congenital homonymous
hemianopia. Am J Ophthalmol 1976;82:498-500.
7. Wellings PC. Detection and recognition of visual field defects
resulting from lesions involving the visual pathways. Aust N Z
] Ophthalmol 1989;17:331-5.
8. McFadzean R, Brosnahan D, Hadley D, Mutlukan E. Repre­
sentation of the visual field in the occipital striate cortex. Br J
Ophthalmol 1994;78:185-90.

MISSING TEMPORAL CRESCENT

349