16 November 2026

Brine Shrimp Phototaxis — Can Tiny Animals Choose Where to Go?

 


Brine Shrimp Phototaxis — Can Tiny Animals Choose Where to Go?

Give a group of tiny swimming animals a light at one end of their container. Will they gather there? And what happens when you move the light?

At first, the water appears to contain little more than orange specks. Look closely, however, and the specks are swimming.

Place a gentle light at one end of the container and watch. Over the next few minutes, the distribution may change. Move the light to the opposite end and you have a wonderfully simple question:

Will the animals follow it?

This investigation explores phototaxis: movement towards or away from a light source. It can turn a small container of brine shrimp into an engaging lesson about behaviour, experimental design and the difference between what we observe and what we think it means.

Allow one to two days for hatching, followed by a short observation session. A full investigation with controls and repeats takes longer than the initial demonstration.

Tiny Animals, Real Biology

Brine shrimp belong to the genus Artemia. They are small crustaceans associated with salty aquatic habitats.

The dried material sold for hatching consists of cysts containing dormant embryos, rather than ordinary empty shells waiting to become animals.

Once conditions are suitable, the embryos resume development and hatch into swimming larvae called nauplii. Newly hatched nauplii are roughly 0.4–0.5 mm long: large enough to see as moving specks, but much more revealing under a hand lens or low-power microscope. They have a light-sensitive naupliar eye.

A diagram gives students a labelled animal. A living culture gives them something more challenging: an animal doing something they have to explain.

Source: FAO: Introduction, biology and ecology of Artemia.

What Does “Phototaxis” Mean?

The word becomes easier when we divide it:

  • Photo- refers to light.
  • Taxis is directed movement in relation to a stimulus.

Positive phototaxis means movement towards light. Negative phototaxis means movement away from it.

Neither should be confused with phototropism, the directional growth response familiar from plant shoots bending towards light.

Nor is every increase in swimming activity phototaxis. An animal might swim faster when illuminated without consistently heading towards or away from the source. A change in activity without a directional response is a different kind of observation.

Young Artemia can show light-seeking behaviour, but their response is not fixed under every condition. Light intensity, previous illumination and other environmental conditions can affect the result. Published work also describes complex responses when illumination is switched on and off.

Our experiment asks what these animals do under these conditions. It does not assume that brine shrimp always swim towards any light.

Research background: Bai and colleagues, 2022.

Are They Really “Making a Decision”?

For children, “Which end will they choose?” is an excellent opening question.

Scientifically, we need to be more precise.

If more shrimp gather near a light, we have evidence of a behavioural response associated with illumination. We have not demonstrated that they consciously compare alternatives in the way a person might.

I like this distinction as a teaching opportunity. We can keep the sense of wonder while improving the language:

“They like the lamp” becomes “More shrimp were counted near the illuminated end.”

The second statement is less dramatic, but much easier to test.

At Hemel Private Tuition, this is the kind of practical I would use to connect curiosity with evidence: make a prediction, observe carefully, then decide how much the result actually allows us to claim.

Hatch First: A Small Culture Is Enough

You do not need thousands of animals in the experimental chamber. A modest, countable group is more useful and easier to care for.

Hatching equipment

  • Artemia cysts from a reputable supplier.
  • A clean hatching vessel.
  • Dechlorinated water and an appropriate aquarium marine salt mix, or the supplier’s hatching medium.
  • Balance or measuring equipment.
  • Thermometer.
  • Aquarium air pump and airline, arranged according to the hatchery instructions.
  • Suitable illumination and a stable warm location.
  • A separate prepared culture vessel for care after the experiment.

Follow the cyst supplier’s instructions. As a practical starting point, many hatchery methods use approximately 25–30 g of suitable salt mix per litre, with temperatures around 25–28 degrees C. The appropriate conditions vary with the cyst source and product.

For a medium specified at 30 g per litre:

Salt required = concentration x final volume

Salt required = 30 g per litre x 0.5 litres = 15 g

Dissolve the salt and make up to a final volume of 500 ml. A ready-made hatching mix may contain additional ingredients, so use its own directions rather than this calculation.

Provide aeration and stable conditions. Expect hatching over roughly 24–48 hours, sometimes sooner in favourable conditions and more slowly in cooler conditions. Record the actual temperature and hatching time.

The FAO hatchery manual describes the importance of temperature, salinity, oxygen and illumination. Its production methods are more elaborate than this small teaching culture; there is no need to add chemical disinfection or decapsulation procedures to this activity.

Source: FAO: Use of Artemia cysts.

Build a Chamber That Allows a Fair Test

Use an elongated, transparent, flat-bottomed container with a modest depth of culture water. A small rectangular observation chamber is ideal.

Mark three equal zones beneath the container:

ZonePosition
ALeft-hand third
BMiddle third
CRight-hand third

These are markings, not internal barriers. The shrimp must be able to swim freely between zones.

You will also need:

  • A low-heat visible LED lamp or torch on a stable support.
  • Black card to reduce stray light and shield reflections.
  • A timer.
  • A wide-bore pipette.
  • A phone or camera mounted above the chamber.
  • A temperature check before and after the trial.

Position the lamp outside the water, shining through one end. Keep the chamber level and avoid direct sunlight.

Transfer a manageable group, perhaps 20–30 nauplii, using a wide-bore pipette and water from their culture. Keep transfer gentle and minimise temperature differences.

Test the camera beforehand. A beautiful setup is not much use if individual animals disappear against the background.

The First Investigation: Will Their Distribution Change?

Before switching on the directional light, invite predictions:

Will most shrimp collect near the light, move away, or remain spread out?

Ask students to give a reason. Different predictions are useful because they give us something to compare with the evidence.

A practical sequence

  1. Allow the animals to settle. Start with a few minutes under weak, diffuse background illumination, without a strong light at either end. Check that transfer currents have stopped.
  2. Record the starting distribution. Photograph the whole chamber and count the shrimp in A, B and C. Do not assume they start evenly distributed.
  3. Illuminate one end. Keep the chamber, camera and background lighting stationary.
  4. Record at fixed times. For example, photograph at 30, 60 and 120 seconds.
  5. Count from the photographs. A still image is much easier to count than a moving population.
  6. Return to the baseline lighting. Allow recovery and check that the distribution and activity are reasonably comparable before another trial.
  7. Reverse the light. Illuminate the opposite end and repeat.

Those timings are a suggested pilot protocol, not guaranteed response times. Adjust them after an initial observation and then keep them consistent.

If shrimp gather near the light and later gather near its new position, that is stronger evidence than a single photograph of a crowded end.

The Control: What Happens Without a Directional Light?

Run a comparison with the same chamber, timing and weak background illumination, but without the end lamp.

This checks whether the animals tend to collect at one end anyway.

Possible reasons include a slight slope, stray window light, reflections or water movement. A no-directional-light control is especially useful when the starting distribution is already uneven.

Alternate which end is illuminated first across replicate groups. Otherwise, “left first” may become tangled with the effects of time, recovery or previous exposure.

Do not shake the container to redistribute the animals. Allow recovery, or use another small group kept under comparable conditions.

Turn “Lots Over There” into Data

Here is an invented example to illustrate recording and calculation. These are not results from an experiment.

Time after light switched onA: illuminated endB: middleC: opposite endTotal
0 seconds9111030
30 seconds149730
60 seconds197430
120 seconds225330

Percentage in illuminated zone = (number in illuminated zone / total number counted) x 100

At 120 seconds:

Percentage in illuminated zone = (22 / 30) x 100 = 73.3%

Because there are three equal zones, a perfectly even distribution would place about 33.3% in each zone. Real small groups fluctuate, so use the starting distribution and control results as well as this reference value.

Define a counting rule before starting: for example, assign an animal to the zone containing the centre of its body. Record unclear counts rather than quietly changing them to fit an expectation.

Plot time on the horizontal axis and percentage in the illuminated zone on the vertical axis.

A graph helps students see whether redistribution is rapid, gradual, inconsistent or absent.

Four Variables Worth Investigating

Change one factor at a time, while keeping water conditions, chamber dimensions, animal age, observation times and group size as consistent as possible.

VariablePractical approachMain complication
Light intensityUse one white LED with several brightness settingsDimming may also alter the lamp’s output characteristics
Wavelength or colourCompare visible red, green and blue LEDs in the same geometryDifferent colours may deliver different light intensities
DirectionIlluminate left, then right, with recovery between trialsPrevious exposure and fixed order can influence the response
TimeCount at agreed intervals after illumination beginsRepeated observations come from the same group

1. Does brighter light produce a stronger response?

Keep the lamp in one position and use its brightness settings. If that is impossible, changing distance is an accessible exploratory method, but it also changes the illumination pattern.

Measure light at a consistent position if suitable equipment is available.

Do not assume that twice the brightness produces twice the response. Look for the relationship supported by your results.

2. Does colour matter?

This is an excellent extension, but a difficult fair test.

A blue LED and a red LED may differ in optical output as well as colour. Coloured filters also transmit different amounts of light.

A lux meter measures brightness weighted towards human vision. Equal lux readings across different colours do not guarantee equal radiant power or equal photon flux. A rigorous wavelength comparison requires suitable calibrated optical measurements and a clearly defined basis for matching illumination.

For a home investigation, report the limitation honestly:

“The response differed between these lamps, but their light outputs were not fully matched.”

Avoid ultraviolet lamps and lasers. Visible LEDs provide plenty to investigate.

3. Will they follow a reversed stimulus?

This is probably the clearest extension for younger children.

If the accumulation shifts after the light changes ends, students can begin to distinguish a response to the stimulus from a preference for a particular physical location.

4. How quickly does the response develop?

Older students can compare the change from baseline at fixed times.

They should also consider the observation window: a photograph at 30 seconds may tell a different story from one taken after two minutes.

Repeats: Thirty Shrimp Do Not Mean Thirty Experiments

This distinction matters at A level.

Thirty shrimp sharing a chamber also share its temperature, water movements and lighting. They are not thirty independent experimental replicates.

Similarly, four photographs of one group are repeated observations of that group.

For a stronger investigation, repeat with several separately tested small groups, ideally including more than one hatching batch. Three groups can be a useful classroom starting point, although this does not guarantee a statistically decisive result.

Keep each group’s results visible. Calculate a mean response if appropriate, but also show the variation. A neat average should not conceal one trial that behaved very differently.

If Nothing Happens, You Still Have a Useful Experiment

An unclear result is a reason to investigate the method.

Check:

  • Was there already a strong light elsewhere in the room?
  • Were the animals actively swimming before the trial?
  • Had transfer currents stopped?
  • Did the lamp warm one end?
  • Were reflections making a second bright region?
  • Did counts miss animals at the edges?
  • Were age and previous lighting conditions comparable?

Video can help distinguish directed swimming from animals simply being carried by water movement.

Even accumulation near a lamp does not, by itself, reveal every mechanism behind the pattern. Tracking individual paths provides stronger evidence about direction than counting the final distribution alone.

I would much rather a student say, “Our observations are inconclusive because the control also collected at the left,” than claim success from a photograph that does not support it.

That is scientific judgement.

Care for the Animals Before, During and After

Plan ongoing care before hatching.

Use small numbers, short trials and low-heat visible light. Keep the water conditions stable, avoid crowding and stop if activity becomes abnormal or temperature drifts. Keep electrical equipment away from spills and wash hands after handling culture equipment.

Return the shrimp gently to a prepared, appropriately aerated culture after observation.

Newly hatched larvae initially rely on internal reserves, but later stages need suitable food. Follow a reputable Artemia culture guide or supplier’s instructions for microalgal or formulated feed, stocking density and water maintenance. Do not assume a hatching vessel will support them indefinitely.

Do not release animals, cysts or culture water into ponds, rivers or the sea. Arrange a contained culture and follow relevant supplier or institutional guidance for eventual disposal.

The aim is a brief observation of normal behaviour, with care continuing afterwards.

Why This Practical Teaches More Than Phototaxis

For younger children, the activity begins with prediction, counting and watching animals move.

For GCSE students, it develops work on stimuli, responses, variables, controls and graphs.

For A-level students, it opens questions about replication, confounding variables, operational definitions and the limits of conclusions.

Useful discussion questions include:

  1. Why should we record the starting distribution?
  2. What would the reversal test add?
  3. Why might red and blue lamps make an unfair wavelength comparison?
  4. Why are repeated photographs not independent experiments?
  5. What observation would weaken our explanation?

A useful writing task is:

Describe the result, suggest an explanation, and identify one limitation.

Those are three different jobs. Students make progress when they learn to keep them distinct.

A Small Animal Can Ask a Big Scientific Question

The appeal of this experiment is immediate: a light changes position, and a population of tiny animals may change position too.

Its educational value comes from what we do next.

We measure the distribution. We compare it with a control. We reverse the stimulus. We check whether heat, currents or uneven lighting could explain what happened.

The most valuable moment is not simply seeing shrimp gather near a lamp. It is hearing a student ask:

“How do we know it was the light?”

That question turns an attractive demonstration into an investigation.

At Hemel Private Tuition, practical science is a way to build precisely that habit: looking closely, testing explanations and learning to make claims that the evidence can support.

15 November 2026

Your Memory Is Not a Video Recording — Can a Question Change What You Remember?

 


Your Memory Is Not a Video Recording — Can a Question Change What You Remember?

You may remember an event clearly and still be wrong about an important detail.

Two Witnesses, One Incident — Two Different Questions

Imagine a minor collision at a junction. Two people see the same cars make contact. Nobody is injured, and both witnesses believe they have a clear recollection.

Later, one is asked to estimate the cars' speed using the word “contacted”. The other hears “smashed”.

Those words suggest rather different events. Could that difference influence what the witnesses report?

This imagined example introduces a question with consequences far beyond road incidents. It matters when families disagree about a conversation, when we hear someone else's account of an event, and when a student insists: “I knew it when I revised it.”

For OCR A-level Psychology H567, Loftus and Palmer (1974) and Grant et al. (1998) form the classic and contemporary pair under the cognitive theme of memory. They investigate different influences on remembering. [1]

Loftus and Palmer: What Difference Can One Word Make?

In the first experiment, 45 students watched seven traffic-accident films. Different groups received different verbs in a question asking them to estimate speed.

Mean estimates ranged from about 32 mph for “contacted” to 41 mph for “smashed”. The films had not changed; the question wording had. [2]

However, a changed estimate does not, by itself, establish a changed memory. Someone uncertain about speed might use the question's wording to choose an answer: response bias.

In a second experiment, 150 students saw an accident film. Groups received a speed question using “smashed”, one using “hit”, or no speed question. A week later, they were asked whether they had seen broken glass. There was none. [2]

Earlier conditionReported seeing glassPercentage
Smashed16 out of 5032%
Hit7 out of 5014%
No speed question6 out of 5012%

The later reports support a reconstructive interpretation: information supplied after an event can become incorporated into its remembered account. They do not prove that every participant's original memory was permanently overwritten. [2]

Notice, too, that most participants in every group correctly denied seeing glass. The effect was a difference between groups, not universal susceptibility. [3]

Remembering Is an Act of Reconstruction

A useful way to approach reconstructive memory is to ask where a remembered detail came from.

Did I actually see it? Did someone tell me about it afterwards? Have I inferred it because it fits the story?

Suppose a relative says, “Do you remember how furious Uncle David was when he arrived?” Before you have considered the event yourself, the question has supplied an interpretation: he was furious.

Compare that with: “What do you remember about his arrival?”

The second question gives you more room to describe your own recollection. This family example is an illustration of suggestive wording, rather than a result from the accident experiments.

That distinction matters in psychology. An everyday example can help explain a theory without becoming evidence that the original study tested every similar situation.

My teaching priority here is to slow down the jump from an interesting finding to an oversized conclusion. “A question influenced later reports” is a much more defensible claim than “a question can make anybody remember anything”.

When Families Remember Different Versions

Think about a family holiday. One person remembers rain, another remembers sunshine, and a third mainly remembers an argument about parking.

Before deciding that someone is lying, ask whether they are describing the same afternoon, the same part of the journey or the same detail.

People can disagree for several reasons. They may have noticed different things, misunderstood each other at the time, or heard different retellings afterwards. Loftus and Palmer's experiments do not establish which explanation applies to a particular family disagreement.

A useful conversation starter is:

“Which part do you remember seeing or hearing yourself?”

It encourages a distinction between observation and interpretation. It also makes room for uncertainty without turning a difference in recollection into an accusation.

Grant et al.: Why Can Recall Change With the Setting?

Now consider a different problem. You revise a topic at home and feel confident, but struggle to retrieve it during a test.

Grant et al. investigated context-dependent memory: whether matching the conditions of learning and testing helps performance. Forty people read an article in silence or with background noise, then completed short-answer and multiple-choice tests in silence or noise. [4]

The material was a two-page article about psychoimmunology. The noise was a recording of a university cafeteria played through headphones; participants in the silent conditions also wore headphones. One participant's data were excluded, leaving 39 for analysis. [3]

Learning conditionTesting conditionRelationship
SilenceSilenceMatching
SilenceNoiseMismatching
NoiseSilenceMismatching
NoiseNoiseMatching

Performance was better in the matching conditions on both types of test. This supports the importance of environmental cues in retrieving newly learned material. It does not demonstrate that noise generally improves learning. [4]

One possible explanation is that features of the learning environment become associated with the material. Their presence at testing can help retrieval. [3]

Grant studied access to learned information. Loftus and Palmer studied the influence of later wording on reports of an event.

“I Knew It When I Revised It” Deserves a Closer Look

The practical implication Grant and colleagues proposed was straightforward: because examinations are usually quiet, studying in silence may help students match their learning and testing conditions. [4]

That is different from claiming you must sit in the exact examination room to revise successfully. Their manipulation concerned background sound, not a requirement to reproduce every feature of a room.

Nor should context become the automatic explanation for every disappointing test. These studies cannot diagnose why an individual student struggled. The student might not have understood the question, learned the material securely or practised producing an answer independently.

In tuition, I would turn “I knew it yesterday” into a practical investigation:

  1. Close the notes and explain the idea in your own words.
  2. Answer a question that uses the idea in an unfamiliar example.
  3. Mark the answer and identify exactly what was missing.
  4. Revisit the question later, including some practice in quiet conditions.

These are teaching suggestions, not procedures tested by Grant. They help us ask a better question: what can the student retrieve and use without the page supplying the answer?

For this topic, try explaining the difference between response bias and reconstruction without using the word “memory” as the whole explanation.

What Do the Studies Share — and Where Do They Differ?

Both studies belong to the cognitive area: they use performance on tasks to investigate mental processes that cannot be observed directly. Their controlled comparisons help isolate particular influences.

Their questions remain distinct:

  • Loftus and Palmer: Does post-event question wording influence what people report?
  • Grant et al.: Does the relationship between learning and testing contexts influence performance?

Grant is not simply a repeat of Loftus and Palmer with a different task. Together, they broaden the theme from the accuracy of an account to the accessibility of learned information.

For H567 students, that is a productive comparison to explain in your own words. OCR's mapping guide retains both studies in the refreshed H569 qualification, while changing the themed-pair structure; this article's pairing follows H567. [1]

Experimental Controls: Why the Details Matter

Consider why everyone in Grant's study wore headphones. Otherwise, a comparison could involve both background sound and the experience of wearing equipment. Keeping that feature consistent helps narrow the explanation for differences in performance. [3]

The broader lesson is that an experimental control needs a purpose. Rather than writing “everything was controlled”, identify the possible alternative explanation that a particular control reduces.

Similarly, showing comparable material across question conditions makes it easier to investigate wording. Asking one group about a dramatic crash and another about a gentle bump would mix up the event with the question.

Control strengthens the comparison. It does not automatically establish how widely the result applies.

Ecological Validity: Explain the Consequence of the Task

“The study has low ecological validity because it was a laboratory experiment” leaves too much unexplained.

Ecological validity concerns how far findings apply to everyday circumstances. A stronger answer identifies a feature of the study, compares it with a relevant real situation, and explains why the difference could matter.

For example:

Loftus and Palmer's participants watched filmed accidents rather than witnessing an unexpected collision in person. A real witness may experience personal danger and have a different viewpoint, which could affect attention and later reporting. The size of the wording effect therefore cannot automatically be assumed to be the same in real incidents.

This is an evaluation of what the task permits us to conclude, not a claim that the study measured danger or emotional involvement.

For Grant:

Learning an article and answering questions resembles some academic work, making the task relevant to students. However, a brief learning-and-testing session differs from revising several subjects over months. The findings therefore provide a limited basis for predicting performance across a complete examination course.

Avoid making ecological validity an all-or-nothing judgement. A task can resemble one part of everyday life while leaving other important features untested.

Try It: Can Wording Influence a Classroom Report?

Use a harmless staged clip, such as someone gently placing a bag on a desk. Avoid distressing footage or a real incident involving the participants.

With voluntary participation, explain that the activity explores descriptions of an everyday event and that the exact comparison will be discussed afterwards. Participants can opt out without penalty.

  1. Show everyone the same clip once.
  2. Randomly give half the group a question describing the bag as being “placed” on the desk and half one describing it as “slammed”.
  3. Ask everyone to rate the force of the movement from 1, very gentle, to 7, very forceful.
  4. Keep the instructions, viewing conditions and response scale the same.
  5. Compare the groups' mean ratings, then explain the wording manipulation and replay the clip.

Mean rating = total of ratings / number of participants

Treat this as a teaching demonstration, not a replication of Loftus and Palmer. A small class may show little difference, and participants may guess the purpose.

Most importantly, an immediate rating difference would show an influence on reports; it would not establish a lasting alteration of memory.

Student Activity: Build Your Own Comparison Table

Close the article and create a table with these headings. Fill it from recall, then check and improve it in another colour.

HeadingLoftus and PalmerGrant et al.
AimInvestigate wording and event reports.Investigate matching learning and testing contexts.
ProcedureAccident films followed by questions; a separate experiment included a delayed test.Article learning and testing under combinations of silence and noise.
FindingsWording influenced speed estimates and later glass reports.Matching contexts produced better test performance.
ApplicationSupports care over suggestive wording.Suggests practising in conditions relevant to quiet tests.
LimitationFilm viewing differs from witnessing an unexpected real collision.A brief task cannot represent months of revision.

Use these short entries as prompts. Expand them with accurate study details rather than memorising the table as a complete exam answer.

Exam-writing challenge: Choose one limitation and write three connected sentences: identify the feature, explain how everyday circumstances differ, then state what this limits us from concluding.

Conclusion: Be Curious About Your Certainty

Remembering deserves more care than pressing play on an imagined internal recording.

The useful response is neither to dismiss every recollection nor to accept every confident account without question. Ask what was observed, what was suggested afterwards and what conditions supported recall.

As a teacher, I want students to leave this topic with more than two researchers' names. I want them to notice a leading question, explain a control and recognise when a conclusion has travelled further than the evidence allows.

The next time you say, “I remember it clearly”, add one more question: “What gives me confidence that this particular detail is right?”

14 November 2026

Make Your Own DNS Server — And Block Adverts Across a Network

 

Make Your Own DNS Server — And Block Adverts Across a Network

A practical Linux project that makes DNS, networking and security visible.

Open a website and you see a page. What you do not see is the collection of requests that helped assemble it: requests for images, fonts, video, analytics and advertising services.

What if students could watch some of that activity happening — and decide which domain lookups their network should answer?

That is where Pi-hole becomes interesting.

It can block requests to many advertising and tracking domains across devices using it for DNS. But its greatest value in a computing lesson is the opportunity to investigate a service that normally works quietly in the background.

The aim is to build something useful, observe its behaviour and explain the evidence.

1. DNS: the question before the connection

DNS stands for Domain Name System. Among its jobs is finding IP addresses associated with names such as www.example.com.

A browser needs an address to establish a connection to a server. DNS helps it find one. The subsequent web connection is a separate operation. [1]

An address-book analogy is useful: looking up somebody's telephone number and making the call are different actions.

For students, that distinction matters. A DNS server does not normally download the webpage on the browser's behalf.

There is also more to DNS than one name matching one address. An A record provides an IPv4 address; an AAAA record provides an IPv6 address. A name can have multiple records, and answers can change.

The first question I would ask is:

If a computer cannot look up a website's address, does that necessarily mean its internet connection has failed?

That question becomes much easier to answer once students operate their own DNS service.

2. What Pi-hole actually does

Pi-hole provides a local DNS service with filtering. In a typical installation, permitted queries that cannot be answered locally are forwarded to a configured upstream DNS resolver. Pi-hole also caches answers for reuse. [2]

For blocked domains, it supplies a blocking response instead of the normal address. Its default NULL mode returns 0.0.0.0 for an A query or :: for an AAAA query. Other blocking modes exist. These are unspecified addresses, not the address of an alternative advert server. [3]

This is often called DNS sinkholing.

Imagine a fictional webpage whose main content comes from one domain while an advertising script comes from another. If the advertising domain is blocked, the browser cannot obtain its normal address through Pi-hole. The main page may still load.

Pi-hole filters domain lookups. It does not inspect each image or video and decide whether it is an advert.

That explains both its usefulness and its limitations.

3. Why this makes a strong computing project

My preference in practical teaching is to give students something they can change, measure and explain. A working service gives technical vocabulary a purpose.

Pi-hole brings together several ideas relevant to GCSE and A-level Computer Science, including OCR H446:

ConceptPractical question
Client-server systemsWhich device asks the question, and which service answers it?
IP addressingWhy must clients be able to find the DNS server reliably?
DNSWhat information does an A or AAAA query request?
DHCPHow can a device discover which DNS server to use?
CachingWhy might a repeated lookup avoid an upstream request?
SecurityWhat can filtering prevent, and what can bypass it?
TestingWhat evidence would demonstrate that our rule caused a change?

Installing software is the beginning. The deeper learning comes from predicting what should happen and investigating what actually happens.

4. Choose a small, manageable starting point

For this project, use:

  • A Raspberry Pi or another small computer running a currently supported Linux distribution.
  • Reliable power and storage.
  • A network connection; Ethernet is convenient for a service that should remain available.
  • A Windows or Linux computer for testing.
  • Access to the router's settings if you later extend the project across your own network.

Pi-hole does not require a Raspberry Pi. Its prerequisites currently list modest minimum resources: 512 MB RAM and 2 GB free storage, with 4 GB recommended. Check the supported operating systems and versions before preparing the host. [4]

A Linux virtual machine on a Windows computer is another possible teaching arrangement. It needs a network configuration that makes it reachable from the test client, usually a bridged or external network adapter. The host computer and virtual machine must remain running while clients depend on them.

For a first investigation, I would use one test client, then expand after verifying the results. Work on a network you own or have permission to administer.

5. Install Pi-hole and establish a baseline

Give the server a predictable address

Arrange a stable LAN address, for example through a router DHCP reservation. Pi-hole's documentation requires a static address or reservation. [4]

For the examples below, assume:

DeviceIllustrative IPv4 address
Router192.168.1.1
Pi-hole192.168.1.10
Test computer192.168.1.50

These are examples, not settings to copy blindly. Use addresses appropriate to your own network and avoid conflicts.

Use the official installation instructions

Update the host using its operating system's normal process. Then follow the current Pi-hole installation guide. One documented method downloads the installer before running it: [5]

wget -O basic-install.sh https://install.pi-hole.net

Review the downloaded script before running it with administrator privileges:

sudo bash basic-install.sh

During setup, select the appropriate network interface and an upstream resolver. For a short, agreed classroom investigation, enable sufficient logging to see the test queries. Record the final dashboard address and login details.

On a typical default setup, the web interface is available at:

http://192.168.1.10/admin/

Use the address and port reported by your installation if they differ. Set a strong administration password; the current command is: [6]

sudo pihole setpassword

Keep this first installation simple. Use the initial blocklist offered during setup and check that it has loaded. Adding numerous extra lists immediately makes it harder to explain why something was blocked.

Query Pi-hole directly

Before changing the computer's normal DNS settings, explicitly send it a test query.

On Windows:

nslookup example.com 192.168.1.10

On Linux, where dig is installed:

dig @192.168.1.10 example.com A
dig @192.168.1.10 example.com AAAA

Record the answer and check that the request appears in Pi-hole's query log. If it does not, investigate connectivity, the service and local firewall rules before proceeding.

This confirms that Pi-hole can answer an explicitly directed query. It does not yet prove that the browser or the rest of the network uses it.

6. The clearest experiment: block, observe, restore

Rather than starting with an unfamiliar advertising domain, use a harmless, predictable target: example.com.

Step A: Record the unblocked response

Run the direct query above. Save the result and note the query-log entry.

Step B: Add one exact rule

In Pi-hole's domain management interface, add example.com as an exact denied domain. Ensure the rule is enabled and applies to the test client, normally through the default group.

Menu labels can vary between releases. The important choice is an exact domain rule rather than a broad regular expression.

Step C: Repeat the same query

Run the identical command again. In default NULL mode, the answer should now contain 0.0.0.0 for A or :: for AAAA, and the log should identify the query as blocked. [3]

Step D: Remove the rule and retest

Remove the temporary denial, then repeat the query and confirm that normal resolution returns. If an answer persists, allow relevant cache entries to expire and retest.

StagePredicted observation
Before denialNormal DNS answer
Exact denial activeBlocking response and blocked log entry
Temporary denial removedNormal resolution restored after any relevant caching

The table describes expected behaviour, not measurements already collected.

Ask students to explain why the evidence supports a causal link. What stayed constant? What changed? Why is restoring the original result useful?

Also ask whether blocking example.com necessarily blocks every subdomain. An exact rule matches the specified name; it is not automatically a wildcard rule.

7. Let one device use Pi-hole normally

Record the test device's original DNS configuration. In its network settings, set its DNS server to the Pi-hole address, taking account of any separately configured IPv6 DNS servers.

Then try a lookup without specifying a server:

nslookup example.com

Check which server answers. Browse a few agreed websites while watching the query log.

The browser may use a different DNS path from the command-line tool, particularly if secure DNS is configured. Treat the log as evidence: did the browser's requests actually arrive?

If testing fails, restore the original DNS settings — often automatic settings supplied by the router — and reconnect the device. Keep that rollback available throughout the investigation.

8. Extend filtering across the network

Once one client works, configure your router's LAN DHCP settings to advertise Pi-hole as the DNS server. Renew client leases or reconnect devices, then verify requests from several clients. Pi-hole's post-install guide describes this arrangement. [7]

Some routers instead forward client DNS requests themselves. This can work, but Pi-hole may identify the router as the requester rather than each individual device.

Do not assume that changing a router's WAN DNS setting is equivalent to configuring what LAN clients receive. Check the router's own documentation.

If the router cannot advertise a custom DNS server, Pi-hole can provide DHCP. That is a separate configuration task: disable the router's DHCP service before enabling Pi-hole's replacement, and record the original settings. Avoid accidentally operating two competing DHCP servers. [7]

Two details deserve particular attention:

A public secondary DNS server can bypass filtering. Clients do not universally treat the second address as an emergency-only option. For filtered redundancy, use a second working Pi-hole rather than assuming an unfiltered resolver will remain unused.

IPv6 needs consideration too. Clients may receive IPv6 DNS information separately. Ensure the DNS resolvers advertised for both protocols match your intended setup. Looking up an AAAA record through Pi-hole's IPv4 address is valid; the query's transport and the address type being requested are different things.

Check for forwarding loops as well. If the router forwards queries to Pi-hole, Pi-hole should not send those same queries back to the router as its upstream resolver.

9. Watch the requests — but interpret them carefully

Open an agreed webpage and inspect the query log. You may see names associated with content delivery, analytics, advertising and background activity.

Ask:

  • Which requests appear when the page opens?
  • Which were blocked, cached or forwarded?
  • Which names remain unfamiliar?
  • What happens when the browser is closed?

An unfamiliar domain is not automatically malicious. A familiar-looking name is not proof of safety. Background apps and operating system services also generate requests.

Nor is the DNS log a complete browsing history. It records domain queries that reached Pi-hole, not full HTTPS URLs or the contents of pages. Caching can mean a page visit produces no fresh lookup, while background activity can produce lookups without a person opening a page.

For teaching, I would use a dedicated test device and agreed activities. Pi-hole offers privacy settings that hide domains, hide clients or reduce recorded detail further. Avoid publishing identifiable query logs in screenshots. [8]

10. What does “25% blocked” really mean?

Suppose a fictional observation period contains:

Total DNS queries = 8,000
Blocked DNS queries = 2,000

Blocked query percentage = (blocked queries / total queries) x 100
Blocked query percentage = (2,000 / 8,000) x 100 = 25%

That tells us about the share of recorded DNS queries blocked during that period.

It does not establish that 25% of internet traffic was advertising, that 25% of bandwidth was saved or that pages loaded 25% faster.

A small background service can repeatedly retry a blocked name. A video stream can transfer a large amount of data after relatively few lookups.

This makes an excellent data-literacy exercise: what does the counter measure, and what additional measurements would justify a stronger claim?

11. Why some adverts remain

Pi-hole works well when unwanted content uses domains that can be blocked separately from useful content.

When adverts and wanted material share a hostname, domain filtering cannot reliably distinguish between them. YouTube adverts are a familiar example: blocking relevant video-serving domains can also break the videos. [9]

Pi-hole also does not tidy up every empty advert space in a page. Browser extensions can operate within the browser and make decisions that a DNS filter cannot.

Filtering can also be bypassed when a device or application uses another resolver. DNS over HTTPS can send lookups to an external provider rather than the local DNS service; Mozilla documents this interaction with local filtering policies. [10]

VPNs, explicit DNS settings and some privacy services can also change the path. Investigate those settings on the agreed test client instead of assuming the router controls every lookup.

12. Troubleshooting is part of the lesson

SymptomUseful next check
Explicit query to Pi-hole times outCheck its address, service, client reachability and DNS firewall rules.
Explicit queries work, but browsing creates no log entriesCheck browser secure DNS, VPNs, client DNS settings and caching.
Only some devices are filteredCheck DHCP leases, manual settings and separately advertised IPv6 DNS.
A useful feature stops workingReproduce it, inspect recent blocked queries, and test a narrow allow rule.
Websites appear to fail when the Pi-hole host stopsRestore the previous DNS configuration and confirm whether DNS availability caused the failure.

More blocklists do not automatically produce a better setup. A false positive — a useful domain blocked by mistake — can break logins, embedded media or other features.

The response should be specific: identify the relevant domain, test the explanation and make the smallest justified exception.

13. Security includes the service you have just created

Domain filtering can reduce access to some known unwanted or harmful destinations when suitable lists are used. It does not replace software updates, careful account security or endpoint protection.

Keep the operating system and Pi-hole updated, protect administration access and retain a configuration backup. The Pi-hole command reference documents update commands; container installations require their own image-update workflow. [6]

Keep DNS and administration access restricted to the intended local network. Do not forward public internet traffic to DNS port 53 or expose the dashboard for this exercise.

There is also an availability lesson. Once other devices depend on your DNS service, switching it off can disrupt name resolution even while broadband remains connected.

Building infrastructure creates a responsibility to keep it available — or provide a clear way to recover.

14. Take the investigation further

Compare cached and forwarded queries

Repeat a permitted lookup and inspect its status in the log. DNS answers are cached for a time governed by their TTL, or time to live. Caches exist at several levels, including the client and resolver. [1]

Use several observations before claiming a speed improvement. A changing query time alone does not prove which cache was responsible.

Compare devices

Use a Windows computer, a Linux computer and an iPad as clients. Do they reach the same DNS service? What background requests differ? Use anonymised observations rather than judging users by their devices' traffic.

Add a local DNS name

Create a local record for a laboratory web server, using a name within home.arpa, such as labserver.home.arpa. Then test name resolution separately from loading its webpage. Use the server's real LAN address.

Explore recursion with Unbound

For a later project, Pi-hole's official Unbound guide explains how to add a local recursive resolver. That creates a useful comparison between forwarding to an upstream provider and resolving through the DNS hierarchy. It is an extension, not a requirement for the first lesson. [2]

15. The real success is being able to explain it

Adverts provide the initial motivation. The lasting educational value is understanding the service behind the effect.

A student who can explain why one query was blocked, another was forwarded and a third never reached Pi-hole has moved beyond following installation instructions.

They are connecting addressing, protocols, caching, security and evidence.

At Hemel Private Tuition, this is the kind of practical investigation I value: a useful system that gives students reasons to ask better questions. For GCSE and A-level learners, including OCR H446 students, it offers a way to connect networking theory with decisions they can observe and justify.

Build the server. Follow a query. Change one rule. Then explain exactly what happened.

13 November 2026

Can We Make Copper Grow Old? The Chemistry of Artificial Patination

 


Can We Make Copper Grow Old? The Chemistry of Artificial Patination

A new sheet of copper has a warm, reddish metallic shine. An old copper roof may be brown, nearly black or green. A bronze statue can carry several colours across the same surface.

Those changes are chemistry made visible.

The metal has reacted with its surroundings, producing a surface layer called a patina. By changing those surroundings deliberately, we can investigate how patinas develop—and why the same metal can age differently in different places.

For students, this offers a particularly satisfying investigation. It connects oxidation, atmospheric chemistry and experimental design with buildings, sculpture and the difficult decisions involved in conservation.

The challenge is to do more than produce an attractive green surface. Can we explain what changed, compare treatments fairly and recognise what our observations cannot establish?

A Patina Is a Reaction Layer

Paint is applied over a surface. A corrosion patina develops through reactions involving the metal itself.

Copper atoms can lose electrons and enter compounds containing oxygen and other elements. The products accumulate, dissolve, crystallise or change again as conditions vary.

A patina may therefore contain several compounds arranged in layers. Its composition depends on moisture, oxygen, pollutants, deposited salts and the history of wetting and drying.

Some artists deliberately encourage these changes. Architects value copper partly because its appearance develops over time. Conservators study the resulting layers because they can contain evidence about an object's manufacture, treatment and environment.

Why “Copper Turns Green” Is Only the Beginning

Students often meet the explanation that copper reacts with air to form green copper carbonate. That introduces a useful idea, but real outdoor surfaces are more complicated.

Examples of relevant compounds include:

CompoundFormulaRelevance
Copper(I) oxide, or cupriteCu2OOften associated with the early reddish-brown surface layer
Copper(II) oxide, or tenoriteCuOBlack oxide; readily demonstrated by heating copper in a separate practical
Basic copper carbonate, malachiteCu2CO3(OH)2A green carbonate-containing corrosion product
Basic copper sulphate, brochantiteCu4SO4(OH)6An important constituent of many mature outdoor patinas
Basic copper chloride, atacamiteCu2Cl(OH)3One chloride-containing product associated with suitable salt-rich conditions

Research on atmospheric corrosion identifies sulphate-rich and chloride-rich patinas in different environments. Carbonates can also occur; there is no single formula for every green roof or statue. [1, 2]

The table describes possible chemistry. It is not a colour chart that lets us identify an unknown coating by eye.

A blue-green surface is an observation. Calling it a particular copper compound requires further evidence.

The Chemistry Underneath the Colour

One possible initial oxidation reaction is:

4Cu + O2 -> 2Cu2O

For a simplified model of basic copper carbonate formation, we can write:

2Cu + O2 + CO2 + H2O -> Cu2CO3(OH)2

This balanced overall equation summarises a possible net change. It does not mean that an outdoor patina forms in one step, or that carbonate must be the main product.

For A-level students, the electrochemical interpretation goes further. In a simplified model involving copper(II) ions:

Cu -> Cu2+ + 2e-

At other sites, oxygen can accept electrons:

O2 + 2H2O + 4e- -> 4OH-

The electrons move through the metal, while ions move through a thin film of water. Subsequent reactions depend on the local solution chemistry. Real copper corrosion can also involve copper(I) intermediates.

This helps explain why a surface that merely looks damp can support chemical change. It does not need to be submerged in a beaker.

The Investigation: Give Identical Copper Samples Different Environments

Suitable level: GCSE upwards, with electrochemistry extensions for A level.

Timing: About 45–60 minutes to prepare; brief daily observations for 7–14 days. Some treatments may show little change in that period. Longer observation can be valuable.

Main question: How do salt and dilute vinegar affect the development of visible surface changes on copper during repeated wetting and drying?

This is a proposed exploratory method, not a guarantee of a particular colour or a proven recipe for a durable architectural finish.

Materials

  • Twelve equal pieces of uncoated copper sheet, approximately 2 cm x 3 cm, prepared with smooth edges by an adult.
  • Four labelled treatment solutions and separate small measuring syringes or pipettes.
  • Distilled or deionised water.
  • Sodium chloride.
  • Plain white vinegar with its labelled acidity recorded; ideally 5% acidity.
  • Fine abrasive pad, detergent, forceps and eye protection.
  • Separate labelled plastic or glass dishes, a protective tray and ventilated splash covers.
  • Camera or phone, ruler and a white or grey reference card.

Use genuine copper from the same sheet. Modern coins may be plated rather than solid copper, and scrap pieces may carry lacquer, solder or unknown contamination.

Prepare Four Treatments

TreatmentPreparationWhat it helps investigate
A: Water control100 mL distilled waterChanges during water wetting and drying
B: Salt solutionDissolve 1.0 g sodium chloride and make up to 100 mL with waterEffect of adding salt compared with A
C: Dilute vinegarMix 20 mL of 5% acidity vinegar with 80 mL waterEffect of dilute acetic acid compared with A
D: Salt plus dilute vinegarDissolve 1.0 g sodium chloride in 20 mL of the same vinegar; add water to a final volume of 100 mLEffect of salt in the vinegar treatment, compared with C

C and D have approximately 1% acetic acid when made from vinegar labelled 5%. If the vinegar has a different labelled acidity, record it and recognise that the treatment strength differs.

Three samples per treatment provide a first look at repeatability. Keep an additional cleaned copper piece untreated if available, as an ambient-air reference.

Prepare the Copper Consistently

  1. Label each dish, keeping labels and adhesive away from the tested copper face.
  2. Lightly abrade every sample using the same pad, number of strokes and direction.
  3. Wash with detergent, rinse thoroughly with water and finish with a distilled-water rinse.
  4. Let the pieces dry and handle them with forceps or by the edges.
  5. Photograph every sample before treatment.

Surface preparation matters. Fingerprints, grease and different abrasion marks can change wetting and the development of the coating.

Apply a Measured Film

Lay each sample horizontally in its own dish. Apply 0.20 mL of its assigned solution to the upper face, spreading it gently with the pipette tip without scratching the surface. Keep the treated area consistent and record any run-off.

Allow the samples to dry at room temperature under ventilated splash covers. Once daily, photograph them before applying a fresh 0.20 mL of the assigned solution. Continue for 7–14 days.

If a surface remains wet at the next observation, record this before adding more. Drying time is itself useful evidence.

Keep temperature, lighting, sample orientation, solution volume and observation times as consistent as practical. Rotate dish positions systematically to reduce the effect of a warmer or draughtier part of the bench.

Do not rinse between daily treatments. This means salts and other residues can accumulate, so the investigation models repeated deposition as well as repeated wetting. State that limitation when interpreting it.

Practical Safety and Clean-Up

Carry this out as supervised laboratory work. Wear eye protection, protect against sharp copper edges, use small quantities and keep food and drink away.

These preparations need no heating, spraying, ammonia, bleach, peroxide or strong acids. Do not introduce household cleaners into the experiment.

Once the liquids contact copper, they may contain copper compounds. Collect used solutions and rinsings in a labelled waste container; keep contaminated wipes and loose deposits contained. Arrange suitable chemical-waste disposal rather than pouring them into drains, soil or ponds. Avoid touching or brushing dry corrosion powders.

Use new copper samples, never historical objects, valued coins, cookware or drinking-water fittings.

What Might Happen—and Why It May Surprise You

The water control may change slowly. Salt-treated copper may develop localised patches, while vinegar-containing treatments may initially appear cleaner or brighter before later discolouration.

That last possibility is especially useful for teaching.

Acetic acid can dissolve existing oxide or other surface deposits. In the presence of oxygen it can also participate in a chemical environment that promotes further copper corrosion. Salt changes the conductivity and chemistry of the surface film; chloride ions can influence the corrosion products and localised attack.

Acetate-containing copper products may occur in vinegar treatments. Chloride-containing products may occur when salt is present. Neither assignment is established by appearance alone. Conservation guidance also identifies acetic acid and airborne chlorides as possible causes of green corrosion layers. [3]

There may also be white salt crystals left by evaporation. Those are not automatically a copper patina.

The treatment that produces the strongest colour is not necessarily the treatment that produces the most stable coating.

Turn Photographs into Evidence

Use the same camera position, background, lighting and exposure settings wherever possible. Photograph dry samples at the same point in the daily cycle, with the reference card in view.

A simple record can include:

DaySampleColour and patternApproximate changed areaSurface notes
0D1Bright reddish copper0%Fine abrasion marks
3D1Record actual observationEstimate from photographNote crystals, wetness or loose material
7D1Record actual observationEstimate from photographCompare with D2 and D3

For a repeatable estimate, place a digital grid over each photograph and apply the same rule for counting altered squares:

Estimated coverage (%) = altered area / total assessed area x 100

Keep the assessment away from labels and cut edges. Decide beforehand whether brown darkening and blue-green patches will be counted together or recorded separately.

Graph coverage against time for each treatment, showing the three individual samples or a mean with their spread. Be clear that colour coverage is a measure of appearance, not a direct measurement of corrosion rate.

Weighing can be an extension, but mass change is difficult to interpret: it may include retained salts, water and corrosion products, while some copper may have dissolved into liquid. A heavier sample has not necessarily suffered less corrosion.

Can We Tell What the Patina Contains?

Colour, texture and experimental history support hypotheses. They rarely establish a unique composition.

A convincing investigation separates three statements:

  • Observation: A blue-green deposit appeared after repeated salt-and-vinegar treatment.
  • Interpretation: The environment encouraged the formation or accumulation of copper-containing surface products.
  • Unresolved question: Which compounds are present, and in what proportions?

Researchers can investigate those questions using methods such as X-ray diffraction and Raman spectroscopy. Published artificial-patination research uses instrumental analysis alongside photographs and mass measurements. [1]

There is a powerful lesson here for examination answers: describe the evidence accurately, then make a conclusion that the evidence can support.

The Carbonate Connection: An Extension with a Clear Limit

To explore carbonate chemistry, introduce a fifth treatment containing 1.0 g sodium hydrogencarbonate made up to 100 mL with water. Use fresh, identically prepared copper samples and the same application schedule.

This changes both the available ions and the pH, so it does not isolate a single variable. Little visible change over a fortnight would be a valid result.

Adding hydrogencarbonate does not guarantee a malachite coating. For a stronger comparison, examine a labelled malachite specimen alongside the experimental surfaces and discuss what would be needed to prove that the same compound is present.

Do not mix this treatment with vinegar: the acid–hydrogencarbonate reaction would create a different investigation.

From the Laboratory Bench to Roofs and Statues

A roof experiences rain, drying, deposited particles and changing temperatures. A sheltered section may remain damp for a different length of time from an exposed edge. Water running from one material onto another can carry dissolved substances with it.

That makes a weathered building a record of its surroundings, although its colour alone cannot reconstruct that history reliably.

A useful outdoor activity is to photograph an accessible copper feature or bronze sculpture from public ground. Compare sheltered and exposed areas, upward-facing and vertical surfaces, and streaks below projections. Do not touch, sample or clean it.

Ask which differences might arise from water flow, previous coatings, alloy composition or maintenance. A bronze statue is an alloy object, so results from pure copper are a useful model with limits.

Conservation: When Cleaning Can Remove the Evidence

For a new decorative object, an artist may deliberately create a patina. For an old object, the existing surface may carry historical and artistic significance.

Polishing it away can remove original material as well as alter the appearance.

Conservators distinguish between relatively stable, adherent patinas and active corrosion associated with continuing deterioration. Fresh powder, flaking or new erupting spots deserve attention; green colour by itself is not a diagnosis. Chloride-related active corrosion is a particular concern for some archaeological copper alloys. [3]

Historic England explains that a compact natural green patina on bronze can reduce the corrosion rate, while maintenance remains necessary. [4]

Our accelerated coatings have not been shown to offer equivalent protection. Demonstrating durability would require a further investigation into adhesion and behaviour during continued exposure.

The conservation question is therefore richer than “How do we make it shiny?” It is: What should be preserved, what is changing, and what evidence justifies intervention?

Why I Would Use This in My Teaching

What appeals to me about this investigation is that students have to return to it.

They make a prediction, prepare samples, notice an unexpected patch, compare photographs and decide whether their original explanation still fits. The learning continues between the first application and the final observation.

At Hemel Private Tuition, that is the kind of practical thinking I want to encourage. A student can know the definition of oxidation and still need practice deciding whether an experiment supports a particular explanation.

I would be especially interested in the student who asks why two supposedly identical pieces look different. That question opens a useful discussion about surface preparation, uneven drying and the value of repeats.

The most attractive sample may make the best photograph. The awkward, inconsistent sample may generate the best scientific discussion.

Conclusion: A Changing Surface Can Change How We See Chemistry

Artificial patination turns a small piece of copper into a meeting point between chemistry, architecture, art and history.

It shows how surface reactions depend on the environment, how several compounds can contribute to a familiar colour, and why corrosion can produce both valued surfaces and serious deterioration.

The strongest outcome is a carefully supported explanation: what we changed, what we observed and what remains uncertain.

The next time you pass a green roof or a weathered bronze statue, look at the patches, streaks and sheltered corners. There is an experiment unfolding there—one that may have been running for generations.

Can we make copper look old? Often, yes. Understanding what we have made is the more interesting challenge.

For GCSE and A-level chemistry tuition that connects practical investigation with clear explanations and examination skills, contact Hemel Private Tuition.

Sources and Further Reading

  1. Bureš and colleagues, Artificial Patination of Copper and Copper Alloys in Wet Atmosphere with Increased Content of SO2, Coatings (2019). Research context, not a method to reproduce in this practical: https://www.mdpi.com/2079-6412/9/12/837
  2. Smithsonian Museum Conservation Institute, abstract of Vernon and Whitby's research on the mineralogical relationships of copper corrosion products: https://mci.si.edu/node/1202800
  3. Canadian Conservation Institute, Recognizing Active Corrosion: https://www.canada.ca/en/conservation-institute/services/conservation-preservation-publications/canadian-conservation-institute-notes/recognizing-active-corrosion.html
  4. Historic England, Bronze Statuary: https://historicengland.org.uk/advice/technical-advice/buildings/maintain-and-conserve-metals/bronze-statues/

Brine Shrimp Phototaxis — Can Tiny Animals Choose Where to Go?

  Brine Shrimp Phototaxis — Can Tiny Animals Choose Where to Go? Give a group of tiny swimming animals a light at one end of their container...